[Long and Old] Push, Shove, Pull, Capture '97.

I Find Karma (adam@cs.caltech.edu)
Wed, 8 Apr 1998 02:29:01 -0700


Yeah, I know it's long and over a year old -- assembling a team of
"experts" to talk about the then-future of push technology -- but I'm
finding

http://www.wordcasters.com/Eventlist/transcripts/techsig397.htm

interesting enough to include below for archival purposes. Watch the
amount of misinformation and antibits conveyed as the discussion
meanders fecklessly: Calling the organization "WWWC"?! Claiming that IP
Multicast can guarantee reliable delivery semantics with no caveats?
Thinking that the Dynamic HTML of early 97 was actually a properly
conceived idea? Predicting that by this point in 1998 everyone on the
Internet will be using publish-subscribe channels to consume content? I
find it amazing how many people are willing to go on public record
saying things that are patently untrue, overtly-known vaporware, or
unfounded sweeping generalizations or predictions...

Speaking of which, does Tibco really own THE patent on
Publish-Subscribe? A number of clever proprietary publish-subscribe
solutions patents I can believe, but why would the patent office give
them THE publish-subscribe patent when it's such an archetypal way to
write push applications?
-- Adam

[full transcript follows]

PUSH, SHOVE, PULL, CAPTURE '97
Microsoft, CompassWare, TIBCO, InterVox
Tuesday, March 18, 1997 6:00 - 8:30 p.m. EST

Peter Chislett: Thanks for trekking out to 11th Avenue and trying to
make it to this space, which is a little bit hard to get to. I apologize
for that.

Tonight we have a great panel that we put together to talk about Push
and its future state. And I'm trying to figure out what is actually
going to happen in this space, and as a moderator we have Jeff Pulver
helping us out. Before we get to that, I just wanted to say special
thanks to a couple of people.

First of all, Expocon, which is kind of supporting us in this room
tonight. Mal Rose, whose assistance in arranging this space with
Expocon. Peggy Miles, whose assistance putting parts of this together.

Susan Quinn, who is with Wordcasters Corporation, and we actually wanted
to Wordcast this live tonight on the Net, but there was no phone line in
this room. And if we wanted one, they wanted my next child and I don't
have one. (laughter) So that's not going to happen, but what is going to
happen is we're taping it. And the people that tape these rooms have
allowed me to borrow -- actually I had to rent it, but it wasn't a lot
of money -- their recorder. And we're going to have a transcript of this
evening available up on the site we hope next week. Our site is
TechSIG.org -- you can put the w's there if you want.

And other thank you's to the Silicon Alley Reporter, Jason, who's got a
great paper mag, it's really a great newsletter. I find it very useful.
So useful that we're actually giving away five free year subscriptions
to it, if you put your business card back there in the little fishbowl
and we'll pick five people and let you know whether you get a year's
subscription, it's, I think, $90 a year, so it's worth putting your
business card back there.

And the Red Herring, who supplied the magazines that I carried here that
are heavy. (laughter) And Wired magazine who has helped a lot. And we've
actually ripped off their cover on our little thing and they don't know
that, but they might know someday. And @NewYork has helped out a little
bit in this, and thank them. And the people from ExtraActive for some
hosting and some help that they've given us and Jeff Harrington.

I just want to go through this real quick. You're going to read all this
stuff. I'm not going to name all this stuff. But there's, the second
page, there's push resources that you can go to, some current articles
that were just published, Gartner Group's got a great discussion going
-- digital deliverance -- it's some good resources that if you're
interested in this media and where it's going, definitely check that
out.

And there's two list services that I'm aware of, and there's probably
more. But the main one that I've found, and Peggy's running for Intervox
and Broadcast.net is sponsoring, it's where I learn a lot about this
industry and it talks about the video/audio side of push as well, which
I actually call webcast and there's a big debate on the list serve about
all that.

The next is IP Multicast, which is an initiative put together by
Stardust, I think it is, and that's talking about more of the technical
aspects of layering TCP/IP a little bit more like the M-Bone technology
that's out there. And that's a little more technical, but the webcasting
list is, Peggy may talk a little about that later.

And then also the conferences which, the big one, the best one, is the
Spring VON Conference. Unfortunately, it's not in the beautiful city of
New York, but the next one will be, I guess. Jeff can plug that because
he can do whatever he wants. (laughter)

It's in San Francisco, April. The next one is in, I guess, San Francisco
again, and others, the Web Broadcasting '97 is in San Jose. And then we
actually get one that takes place here at the Plaza Hotel and that's in
July. And in the back of the room there are flyers for each one of those
conferences and you can pick one up on your way out.

And then there's three organizations -- I've already mentioned one of
them -- the WWWAC TechSIG, which has kind of been around and we've had
some meetings but we haven't had one as big as this. And that's our Web
site, as I said earlier. And the Internic International Webcasting
Association, which Peggy will probably talk a little bit about, which
she's involved in, she's co- chair, I think the title is. And that's it
for me. I guess I should have brought these up as props and didn't use
them. But I think I'm going to turn this over to Jeff right now because
he's a professional and I'm an amateur at this. So Jeff, take it away.

Jeff Pulver: Thank you very much. Good evening. I wanted to thank Peter
for inviting me to moderate this evening's session on push/pull
technology. Before we get started, if I could just see by a show of
hands, how many people here in the room have heard about and are
familiar with push/pull?

Good. Are there any people who have their own channels up now? Anyone
thinking about launching a channel in the next three months? Next six
months? Just a few, okay.

The next year? All right.

The industry has certainly grown. You talk about Internet time being a
lifetime. In the last six months, really, it went from an industry that
was just a buzz to a space now there are about 35 different companies
who are delivering different types of technology whether it's a push or
a pull. I mean clearly, I think e-mail is an ultimate push. You know you
write an e-mail, you send it to someone, they get it, you don't know
whether they read it, but at least it's being delivered. It doesn't
solve all the paradigms, but at least e-mail to me is still one of the
killer apps on the Net because it just works.

On the other side of the coin is Web pages, which to me is really an
ultimate pull, because you can have a great Web site, but you have to
get people there to pull down your content and take a look at it. Over
the last year, several utility companies came out which were taking
people's content offline at night, products like Freeloader, which would
surf the Net for you and bring down content. It doesn't mean you ever
read it, but it just meant that it brought it down for you
automatically.

Then came PointCast, right? The whole idea of being able to push content
onto your desktop in a corporate environment. The one thing which Chris
Hassett was successful at is he guaranteed within a year he'd reach a
million desktops. And he has. I can't say that a million or so people
are happy seeing all those commercials, but rumors that News Corp. is
going to pay $300 million to buy his company appearing in the Wall
Street Journal can't hurt him either. And it's certainly helped validate
the space.

What we will be discussing here tonight with the distinguished panelists
is a little bit about the technology, at least that's what I hope to get
out of this. And the presentations will be really in four pieces. Talk a
little bit about the core technology, the platform. Then we'll get into
the plumbing. Do a little filtering, and then take a look at things from
a channel implementation perspective.

If you have questions, if it's all right with the panelists, please
raise your hand while they're talking to make this a little interactive,
because certainly the reason we're here today is to talk to you in not
so much in a broadcast fashion, but something which is more interactive.
And if there is something being said that you'd like us to repeat,
please just ask us to do so. And we'll try to repeat the questions so
everybody here can hear them.

Our first speaker is from Microsoft, it's Tal Saraf and he'll be
discussing, I hope, a little bit about the announcement which Microsoft
made last week, discussing the standards for push technology.

Tal Saraf: Thank you. I'm actually from Microsoft's Developer Relations
Group and I guess I should introduce myself briefly. My role in
Microsoft is actually talking to many of the companies that you all
represent here in the New York area. I'm actually based in New York
while my group is in Redmond. And so I spend about half my time talking
to the large content sites, so you can imagine all the big content sites
in New York. And then the other half of the time, really focused on
talking to the companies that would help those people bring up their
sites, either the long-term solution providers or other companies that
would potentially be involved in the hosting process. And what I'd
really like to talk about tonight has to do with some of the
announcements we made at Internet World in Los Angeles last week. One of
which is really the channel definition format. And I'm sure many of you
have started to read about that -- there have been probably six or seven
articles I saw in the last week on that topic.

And just to give you a sense, this is a specification that we submitted
to the WWWC. It was actually accepted as a submission on Monday, this
past Monday, and was submitted over the weekend. And it really builds
off of XML. So for those of you familiar with that, that's sort of a
broad industry initiative that's been underway. And along with that
release, we actually had a press release that talked about all the
various people who have committed to use CDF, and obviously one of those
is actually on stage here tonight as well, so I'm sure David will be
getting into what their initiative is, but there are others.

Another local sort of New York company is AirMedia and they're working
with their live product. BackWeb is another, FirstFloor software, which
will work with e-mail. Then DataChannel, Diffusion, FreeLoader, IFusion,
Lanacom, PointCast, and about 22 content developers were part of that
press release. One of which was AOL and various others who will be using
CDF.

And the whole idea there is really pretty simple: When the question was
asked, "How many of you are looking to create a channel?" -- our feeling
is it should be very simple to create a channel. It should really be
this concept of HTML plus scripting plus controls. Whatever the
scripting language you want to use, whatever the control technology you
want to use, it should just look and act like a Web page. And so that
sort of summarizes what CDF is all about. And we can obviously get more
into that.

For those of you who are interested in terms of what it would look like,
you can actually get it off the Web right now if you go to
www.microsoft.com/standards/cdf.htm is the actual page, but if you look
at the standards there will be other things up there as well.

The other thing that actually just hit the wire in the last probably
half-hour to an hour is something we also announced at Internet World in
L.A., which is dynamic HTML. And dynamic HTML is sort of the next
ability of HTML, and there will be a number of companies in that release
as well -- I don't have the final release with me, otherwise I would
tell you who is in it. I think CompassWare is actually in that one as
well. So I do want to mention that.

But the idea there is many of the vector capabilities that you might be
familiar with if you have ever seen Future Splash or I guess now it's
called just Flash by Metro Media or have been to maybe MSN.com, you have
seen some of the vector graphics. That's kind of built into dynamic HTML
-- being able to provide that level of capability. And we think that's
really important when we start talking about the channel paradigm. It's
going to be a lot more visual and interactive. It's not necessarily
going to require me to click through everything, and I think that's
really important the more we get into webcasting. I think the closer Web
pages and Web sites are going to look to television. And some of the
things we do with NetShow, and I'm sure other people will mention that,
tie in there.

But really, for me as a user sitting in front of a webcasting machine, I
should be able to just look through the pages. I shouldn't have to click
from site to site to site or link to link. And that's really going to
change a lot of how we do things. And I think dynamic HTML will be very
important along with CDF. And I think I'll leave it off there.

Jeff Pulver: Thank you. If I could just before we move to the next
panelist, could you just explain why Microsoft came out with standards
at this time?

Tal Saraf: Sure. I think one of the things we saw, obviously there were
about 30+ companies working on channels, and even for ourselves, we're
working with many of the "channel companies" in terms of getting content
into Internet Explorer 4, and we felt that it was really important to
make it consistent so that many of you as content developers would be
able to write that content once. We saw re-purposing content as being
relatively expensive. Especially if you had to do one channel for
BackWeb and another channel for IFusion and another channel for IE, if
we did something different. So we really wanted to get together sort of
as an industry and say, "Look, it makes a lot more sense and much
cheaper for all of us if there is a single channel sort of standard that
the WWWC owns, the WWWC and that standards organization walks through
the process to make sure that everyone has a fair shot at creating a
channel once and having it used across multiple products and platforms."

>From the Floor: How are you defining "channel"?

Tal Saraf: I'm really thinking of it in a pretty broad manner, because
the channel definition is a little bit different depending on whose
technology you're approaching it from. And if you look at CDF, it's only
like an eight page spec. So there is plenty of room for customization
and enhancement from one product to another. But the base level channel,
I really think of as just being sort of that HTML plus scripting plus
controls. So you should think of it as just a simple Web page, right?
Well, hopefully not just a simple Web page, because it should be
something compelling, otherwise I am never going to go to it. But from a
technology perspective it should look like a Web page.

Jeff Pulver: Just one more question before we move on. What is the
timetable for approval through the WWWC?

Tal Saraf: Yeah, the WWWC, you can never really judge how long it's
going to take to go through any standards body. Dynamic HTML, for
instance, is actually something that the WWWC has been working on for
eight months. And while the WWWC has said, "We're going to do dynamic
HTML" and all the members are starting to accept it and use it, it has
been a fairly involved process, in terms of back and forth on changing
the specifications to meet all the needs of the various members. And I
can't say that CDF would be quicker or slower in that time frame, but
the standards are always evolving. If you look at HTML itself, that's
been years, right?

Jeff Pulver: Sure, but at least 1.0 came out. My real question is will
IE4 ship before such standards are accepted?

Tal Saraf: I would expect the beta of IE4 to be out before specification
is finalized. We will make sure that IE4.0 supports CDF from whatever
version eventually is ratified by the WWWC. I can't say whether or not
there would be some incremental changes after IE4 ships, but
realistically I think they will be pretty much in sync.

Jeff Pulver: Thank you. Our next speaker is Chris Mather from TIBCO who
is going to be talking a little bit about the plumbing. And one of the
vendors in the push space, for example, that is using TIBCO technology
is Intermind, another one is BackWeb.

Just a little bit of background. TIBCO is formerly known as Teknekron.
And they made a lot of their money by providing the technology behind
many of the trading floors -- right here on Wall Street -- providing
realtime trading systems and realtime information systems. And in fact,
their technology is embedded in some other software products that you
may be familiar with.

But the significant amount of technology that went into delivering
realtime information has found itself now in the content push/pull
space. When TIBCO gets involved in a project, you have to take this very
seriously, because they have spent millions of dollars specializing and
focusing on what realtime delivery is rather than real-enough time. And
the ability to deliver it in a hostile, well, sometimes hostile
environment (laughter) is pretty powerful. Anyway, go ahead, Chris.

Chris Martha: Thanks, Jeff. I want to give a quick overview of what
TIBCO is and what we do, because a lot of times the plumbing software
that gets the messaging content from the provider to the desktop or to
the GUI interface is kind of hidden behind the scenes, and it's kind of
just assumed that it's going to get there efficiently and in a scaleable
fashion. And if messages are lost, they'll be retransmitted. And
basically what TIBCO has been doing for the last ten years, like Jeff
was saying, is providing the plumbing for the middleware, the messaging
middleware, that gets content from the provider or the publisher to the
end user or the subscriber.

So we have this patented technology we call Publish-Subscribe that
allows a user or an interested party to subscribe to some information
one time, and they will continue to get that information -- they will
get that information initially and they will continue to get updates in
realtime as events occur. So we've coined what we call "Event Driven
Computing." And this is where we feel we are very strong in the plumbing
market or in the middleware market, in that we have optimized and we
have worked very hard at making a scaleable middleware so that content
providers can publish information. And they can publish, for example, a
single message once that many, many users, thousands of users, for
example, are interested in. And that single message goes to all users at
the same time.

So we've been on the trading floors, where Jeff was describing sometimes
can be a very tense--

Jeff Pulver: Hostile.

Chris Martha: Hostile environment. And basically your stuff has to work
or it's not going to last. And that's why we're pretty happy with what
we have now, we're calling this "Industrial Strength Technology,"
because it's been proven for about ten years on the trading floors where
the stuff has to work. It has to be load-balancing. It has to be
scaleable. It has to be fault- tolerant. If processes die, there has to
be something as a hot standby ready to pick up so the user doesn't miss
a tick.

And this is really where we've proven our technology and we figured the
next step, now that the Internet is becoming potentially a push
computing, or a push information delivery world, we think we have the
solution for the plumbing. And what we're kind of excited about and
happy about is that all of the other panelists, we feel we have a
compliment to their products. And the ones Jeff was mentioning --
BackWeb, Intermind, Diffusion, DataChannel -- a lot of these companies
have a product that can use this middleware that we provide -- or this
plumbing.

And essentially we feel that to get a true push environment on the
Internet, you're going to need an event-driven dynamic environment that
doesn't require pulling behind the scenes.

So there is a lot of talk and a lot of papers on push computing. And in
many cases what's happening behind the scenes is, even though a client
or user doesn't officially press a button to re-request something,
something is happening behind the scenes that's pulling back to a server
and it's more of a client/server environment still.

And what we are providing is what I'm calling "true push," which is
there is a provider and there are many subscribers and the subscribers
don't have to pull. They subscribe to the information one time, and
whenever an event occurs, whenever an update needs to be published, the
publisher publishes it, and basically in realtime all subscribers get it
at the same time.

>From the Floor: What's the difference between the two standards? TIBCO
in November or December went to the WWWC with a standard?

Chris Martha: Yeah, with the TIBnet.

>From the Floor: Could you explain real quickly what the differences are?

Chris Martha: Okay. Really quickly. The TIBnet that Peter's talking
about is really an extension to the original TIB in that to make a true
push model work over the Internet, you need the infrastructure to allow
that to occur. There are actually many pieces to allow a publisher of
some content to publish out a message. And if you have a million users
spread out over the globe that are interested in that, that message
needs to be filtered and distributed to all of those users through the
infrastructure. So what needs to happen in the Internet is you can't
have point-to-point connections anymore for each one of those million.
You want one message to go out and you want it to go to everybody.

Some of you are aware of the M-Bone, which is a multi-cast backbone
which allows for one publisher to publish. And then everybody connected
to that M- Bone has essentially opened up the routers to allow that
content to go through, and it goes through on what's called a multi-cast
address.

What the TIBnet initiative and the TIBnet standard is -- its vision is
-- is to enable all of the routers on the Internet to be TIB-enabled or
to be TIB- intelligent in that when a message goes out, it goes out on
what's called a subject name. All of the routers will know if any user
that's within their domain wants that data or wants that content.

Jeff Pulver: They have done that because they have subscribed to the
information.

Chris Martha: They have done that because they've subscribed. So a user
will subscribe essentially to a subject and that subject could be a
stock quote or whatever. The router that's connecting that domain to the
Internet is going to be aware of that subject and whenever an update for
that subject comes down the line that will get forwarded down to that
user or to any number of users in that domain.

What's important about this is that in the current M-Bone, the way
multi-cast works is you have one multi-cast address and there is no more
filtering beyond that. So whatever comes down that pipe you get. So it
could be the Rolling Stones Concert or it could be stock quotes, or it
could be whatever.

What we're adding with the TIBnet is what we call subject name
filtering. So you are going to have subject names enabled within the
routers so that they'll be intelligent enough to -- when they get a
message, they'll send it down to their users. And also if their user
subscribes to something and they know that they need to forward that
subject to another set of routers out on the Internet. They will forward
that subject and say, "Please give me this content if it comes to you."

Jeff Pulver: Howard.

>From the floor - Howard: Hi, thanks. I haven't been following your
standards so I'm just asking an infrastructure question, which is are
you planning on posing a separate infrastructure like the M-Bone is now
a separate infrastructure? And if so, what will fund or what will
incentivize the ISPs of the world to build a separate multi-casting
network that would be for data?

I worked with Teknekron previously at J.P. Morgan, so I understand how
there is a separate infrastructure typically for the messaging network.
And I understand the reason for that is the trading environment. I'm
just trying to figure out how it would work in an Internet environment.

Chris Martha: Okay. Another good question. The answer is we are
currently working with ISPs right now that have essentially enabled an
experiment of the TIBnet. They have enabled IP Multi-cast with the
TIBnet architecture to find. And essentially you do not have to have a
separate complete architecture running for standard Internet and for
TIBnet. TIBnet will run within the ISPs. They will essentially turn on
IP Multi-cast -- TIBnet-enabled IP Multi-cast. And anything that was
currently running before that will still continue to work fine,
point-to-point or whatever it is. And if you want to get TIBnet content
then essentially you will connect to the IP Multi-cast address that's
being used. If you don't want any TIBnet content you won't get any
because you won't subscribe to any of the channels.

But the ISPs, the incentive -- the basic thinking behind what we are
providing is we really want to reduce traffic. We're trying to make
everything more efficient. We're trying to actually reduce the traffic,
so it goes back to the whole concept, if you have a thousand users
interested in one piece of information and they have to go to a Web site
today to get it, that means one Web server is handing the same piece of
information one thousand times, over and over and over again. Whereas
with the TIBnet, they send it once and all thousand get it at the same
time.

Jeff Pulver: Our next speaker will be David Waxman. If you're just
following what we're doing here, we're doing the platform, going over
the plumbing. David's company, CompassWare, does may things, but David
is going to be talking about filtering of the information.

David Waxman: We're not necessarily a push company, or a pull company
for that matter, or a push/pull company, just in case you're a little
confused there, or a shove company for that matter, or a capture
company. What we are is a filtering company, a filtering and a relevance
company. And what that means is that -- we've been talking about the
standards that are coming into place that will enable virtually anybody
who has created a Web page in the past to turn that Web page into a
channel. We've talked about the infrastructure that is actually going to
enable that content to be delivered in realtime.

And what that adds up to is a tremendous amount of -- and you'll forgive
the use of the non-technical term here, but I come from a non-technical
background -- a whole load of stuff coming at people. And this stuff is
really very valuable. There is going to be all sorts of different
content providers, ranging from news providers, ranging from people who
want to sell you things, ranging from people who want to tell them about
yourselves, ranging from people who want to create content about
technical information that might be of value, about markets, about
products. And it's all going to be directed at companies that have
servers, that have IP addresses that are publicly available. It's going
to be directed at people that are subscribing to specific channels, or
what in the past have been known as Web pages, that are being delivered
using a standard that's going to be available virtually to everybody,
and using plumbing that's going to enable that information to be
delivered in realtime.

And there's going to have to be some companies that can ensure that the
users who are getting all this stuff can mine it for the valuable
content. Because there is going to be -- think about the Web today. What
percentage of that content is actually valuable to you, do you need?
Probably less than half a percent. And just picture a big chunk of that,
or even a small subset of that, coming at you live, directed at your
server or at your desktop. What that means is information overload.

So what we've done is taken a core technology that we've developed
called Magna technology. Magna technology is a very powerful free-text
analysis, filtering, and search technology. And what we've done is we
built an application that is server-based called InfoMagnet that is
designed to filter all the stuff that's going to be coming at server [ ]
enterprises as well as people's desktops.

And InfoMagnet is an application that as I mentioned has been designed
to -- initially from its genesis as a news management tool -- to filter
realtime news applications that perhaps would have been delivered by the
Teknekron folks in a financial world a couple of years back. Of course,
what we're talking about today, all the Web broadcast channels. And
simultaneously, and this is obviously not of as much interest in
relation to the push model we're talking about, but also simultaneously
search and pull relevant information from all sorts of different static
information. I realize that's not en vogue these days, but still there
is a significant amount of static information out there on corporate
networks that's of value, sitting on file servers, sitting on Web
servers, good old Web sites as well, and of course information sitting
in Lotus Notes databases, you know, legacy data, that sort of stuff, and
relational databases as well.

And what InfoMagnet does is it aggregates all that different relevant
content based on users' profiles, and then just presents the relevant
information to individual users on their own personal Web page or alerts
them based on what they are interested in.

And the critical element of that is our ability to analyze all the text
coming at the user, be it dynamically broadcast over a Web-based
channel, a realtime news feed, or perhaps sitting in a statically -- in
an information base, or a Lotus Notes database, or sitting on a file
server, and ensure that the information is relevant.

And without applications like these driven by core technologies like
Magna technology, everything that we're talking about is not going to
really have the optimal value to the user. Because if hundreds of
thousands of publishers are actually going out and publishing on a CDF
standard, and hundreds of thousands of publishers are shipping their
information with TIBCO's technology or technology like that, unless they
have technology like CompassWare's filtering it at the server, on the
fly, only letting relevant information into the corporate network that
is needed by the users, there is going to be a tremendous backlash
against the whole push industry. Because before you know it, people are
going to be saying, "I don't need all this crap on my computer. I don't
need all this garbage in my network. I don't need my bandwidth to be
taken up. I don't need my storage space requirements to go up
exponentially because a lot of this content is very rich, it's
multimedia. I only want the most relevant information." Otherwise there
would be a significant backlash in the push industry.

And I think where we sit in this whole amalgamation of companies is
actually neutral to all the different content providers, the different
infrastructure providers, the different standards providers, is that
what we do is work with everybody and ensure we add value to pretty much
everything that everybody is doing.

>From the Floor: We are dealing with a push, how are you selecting what
you want?

David Waxman: Well, you're selecting what you want, but on a very, very
broad and general sense. You're selecting a channel. Now let's take an
analogy. When you watch TV, let's say you're watching a movie, or better
yet you're watching the news, that's much more relevant an analogy. You
turn on NBC News at 7:00 on an evening. And maybe there is, out of 30
minutes of broadcast content, 2 minutes that's interesting to you and
the rest of the time you're tuning out. Now when you have your TV on in
your home, there is no cost, a little background noise. But if you have
your channel on the entire time coming into your intranet server and you
are an MIS Manager, there is a tremendous cost to bringing all of that
information in when you only want 2 of that 30 minutes.

>From the Floor: If you are only filtering a broad selection, then you
will continue to get far more data than you want. Wouldn't the MIS
Managers

David Waxman: Not if the filtering technology is sensitive enough to
your needs and actually learns from you automatically over time and
applies all sorts of algorithms. Right now the need is not as
universally recognized because people are used to dealing with tagged
data. For instance, PointCast, for example, let's say their editors tag
a story being about software. But if you're an equity analyst at an
investment bank, you don't want to read all stories about software, or
even all stories about software in the client/server space. You want to
read all stories about software in the client/server space related to
mergers and acquisitions over $50 million. You don't have time to read a
lot of other information. And as the amount of information exponentially
increases, the need for even non-mission critical users, like all of us
here perhaps, are going to need that type of true granularity and
filtering capability to get beyond what the editors think they want to
see. And that's based on the ability to analyze the full text of the
content.

>From the Floor: [inaudible question]

David Waxman: Well, that's actually something we're not addressing.
There might be other vendors that emerge that do address that. And in
fact, I recall there were some rogue products around a year and a half
or so ago -- you'll forgive the term, and if you want to bleep me on the
tape, feel free -- but there was a small company called NoShit Software
out there (laughter) that actually had the capability to block out ads
downloaded on Web pages. But that's really something that was not
accepted by the industry. And that's not something we're looking to do,
because the industry is supported by advertising, the content industry
is. What we're doing is filtering based on the full text, not based on
the actual type of content, whether it's an ad or actual content itself.
We're not doing something like that.

>From the Floor: [inaudible question]

Jeff Pulver: This space is getting very crowded with Cognisoft being
acquired by Varity, and Firefly and all the agent technology in the back
ending married with push/pull technology. I see it as really very
valuable. I can see myself on a desktop, I don't want to see ads. And I
want to have that filtered out if I'm inside a company. And that's what
push/pull is headed towards, is in the intranet and how it's going to be
controlled rather than kind of a rogue set of ad-supported channels. The
Gardner Group has a great discussion going about ads and about these
corporations that really don't, a lot of them don't even want them
anywhere near the desktop. And I think the technology is going to be
supported by licensing fees and by purchasing rather than advertising in
the intranet space. But the Internet space is a whole 'nother ball game.

David Waxman: We agree. The intranet space, that's why we have a
business model that's perhaps somewhat different than some of the other
companies out there. The feedback we've had from people who hold the
purse strings to buy server-based products is that they don't want to
have ads floating through their network. And since they control what's
coming in on a server basis, they have that right to filter it out.

And if you look at the early days of PointCast, there were actually a
couple of issues there that I think we could take lessons from. One is
when everybody was, as Peter says, in a rogue sort of fashion,
downloading PointCast Client and bringing all that extra content into
the enterprise, it was bogging networks down. And the feedback that
PointCast had from network managers was, "If you want to allow this
stuff into the enterprise, create a server-based product that actually
starts to limit the amount of bandwidth that's taken up by the network."

Some other feedback they had was, "Hey, we don't necessarily want our
users looking at ads." So what we're doing is partnering up with all
sorts of different channel providers that are not only providing
external content with ads, but also dealing with all the internal
information which really is a value, finding the relevant internal
information, which is really a value, or the internal channels, the
corporate announcements, that's a real value to the MIS managers that
are buying products to ensure that the users are getting relevant
information. And that's kind of our philosophy on things.

David Waxman: There were a number of companies that actually did not
want to have advertising from vendors. They actually wanted to have
advertising internally. Because the HR department or some department
actually wanted to flash out memos and stuff through the analogy of
banner ads, in this case pushing it through to everyone who subscribed.
So where commercial advertising may be banned, I do think you'll see a
trend of internal departments positioning to push messages through that
same metaphor.

One of the other things I noticed with PointCast was the fact that, in
my former life I was a VP of information technology on Wall Street. And
looking at our networks, one of the things that we got really annoyed
about was that if there were 100 people subscribed to PointCast, we get
it back in the morning. And people have an icon that there was a new
version of software available, they click on an icon. It wasn't to
actually download it, it was actually downloaded for you, and all you
had to do was click on it to be installed.

Now that might be okay for some people, but when you're running a huge
trading operation and all of a sudden this traffic was out of your
control, starts coming down your net, it doesn't make you feel very warm
and fuzzy. What PointCast effectively did is to use HTTP to pierce our
firewall and to allow - - as long it's all HTTP, you could put content
through no matter what it was, right? So they're actually able to, on a
rogue basis, download new clients and take the information control away
from the IT Manager which, while it was pretty effective, it was also
very scary.

Tal Saraf: So the gentleman's point in the back is really an important
one, which is part of this information really has to be relevant to MIS,
right? And so certainly we're working with some of the content companies
to make sure that there are interesting channels such that not all the
IS Managers shut down channels, right? There has to be some business
value in these channels. And we're very, very interested in that. Just
to make sure that there is compelling enough content so that the same
thing that happened with PointCast initially doesn't happen now. And I
would definitely agree, I've already seen internally, many of our sites
now have advertisements for other sites on our intranet or various
things I need to know that get pushed to me today telling me about the
401k application, or HR training, or whatever the internal application
happens to be. So we are starting to see that already.

David Waxman: It replaces the bulletin board by the cafeteria with
bulletins on your Web site and intranet sites.

Jeff Pulver: Hopefully this filtering device that you're talking about
will make the Web more interesting rather than a lot of the glorified
catalogue nonsense that we seem to be seeing the Internet going towards
with a lot of corporate banners and stuff like that. I'm just wondering
what the future is going to be, what you see for your next products in
the developments in there. After you get your filtering device, what do
you come up with next?

David Waxman: Well, actually this is just the first step to a whole set
of intranet-based products. Filtering is just the first step. Once you
actually have the information that you need, then the next thing you
want to do is manage it, collaborate it, and transact with it. And we're
partnering up with document management vendors, we're partnering up with
Internet commerce vendors to make that happen by not only providing
intranet applications, but bringing in other technologies into our
InfoMagnet envelope and extending and embracing their technologies,
extending the functionality to do some of those other things as well.
Because finding and filtering is the first step, but it's the first
critical step.

Jeff Pulver: I see that there's sort of three issues, of which two are
really being discussed. One is realtime, and the filtering is really the
breadth. What about timing? One of the things that made the VCR an
interesting toy is that you could control your timing. And that whole
dimension is somehow not coming up in any of this.

Tal Saraf: I should have been clearer about that. Part of CDF is this
concept, I can either be online and going after this content or I can
basically prepare myself to browse offline. So one of the things that
you'll be able to do in IE4.0 -- and other browsers, too, I'm sure -- is
this concept of being able to surf offline a la FreeLoader or some of
the other packages that are already out today. Which is, go out, collect
the information from these channels, whether they are these sort of
these new style CDF channels that are letting me personalize the content
such that I only get what I care about, or to be able to crawl older
sites. So today I could subscribe to sort of any channel I want, but
it's an older style channel that I crawl. And when I do that I would be
able to surf offline, and I can specify when to go there, how often to
update it, what type of changes I want, how many layers I want to go in,
how much space I want to accumulate on my hard drive for a particular
site, whether I'll follow them off the site, and so forth.

Jeff Pulver: Before we get to Peggy, and she's going to talk about some
audio and video stuff which some people out here I think want to hear a
little bit about, I need to make an announcement that I didn't make
earlier. The WWWAC Organization is having their next meeting March 27,
and the location is to be announced. But Laura Stein, co-worker of Tal
and Stan, is with Microsoft, and she, not too long ago, there was a big
discussion about M3P initiative here in content deals that were going to
be made. She's going to be talking on the panel with an entertainment
lawyer who does Hollywood content deals. So that's Thursday, March 27.

And tomorrow night in this room the BizSig has the room, and they can
fill it like we did. And it's called "Meet The Money. What are Angels,
Strategic Investors, and VC's looking for today? Is it the people, the
process, or the technology?" Bert Alamanski, Jerry Kolana from Flatarm
Partners, Bobby Orbock, and Mike Tannen are speakers. A little
commercial message in between all this. Thanks.

Chris Waxman: A real push.

>From the Floor: I was wondering if you could just clarify how far the
CDF standard goes. I mean, for example, does it mean that we don't have
to have the BackWeb client or the PointCast client, that we have a
generic way to request channels and we don't need those? And all of the
servers from those providers will feed us?

Tal Saraf: You actually hit what I was going to say, which is you don't
necessarily need a particular client or a particular server. What
basically happens is any client could talk to any server.

Jeff Pulver: It's inter-operable.

Tal Saraf: Right. Any client to any server, any server to any client. So
if you're just going to go out and put out a request for, let's say
PointCast content, let's say AOL content. Right? And you would get it
regardless of the client involved.

There are actually three sort of pieces to CDF. One is automatic, so
you'd be able to schedule or dynamically go out and get content. The
other is talk to any server. And the third is compatibility on the
browser side. And I probably should have expanded upon that. When I went
through CDF, I was trying to be brief.

>From the Floor: Just a quick question. From a competitive perspective,
what does that mean to the proprietary push companies, the BackWebs,
Marimbas, those sorts of folks?

Tal Saraf: Okay, for the standard client, this is an area for them to
enhance their clients and not only get all the content that all the
other clients would get, but customize it in one way or another, maybe
by bundling in a filtering package, maybe by having a better UI.

This would not include probably something like a Marimba. Today Marimba
is very Java-focused, and so they're not looking at HTML or XML, but
really a Java-based applet. So while we did talk to Marimba before
releasing this spec and we're very interested in working with them, it's
fair to say that they weren't really interested at this stage.

Jeff Pulver: If you look at it as Microsoft providing ubiquity to the
environment, so it really doesn't matter what the client and what the
server is, as long as everyone adheres to the standard, then people can
then inter- operate with each other. And it's like every other industry
where you have ubiquity, people differentiate themselves by adding
features and enhancements to their own product sets.

David Waxman: Yeah, I think this is great for the industry, because it's
good, it's going to drive the creation of filtering vendors like us,
assuming they have the filtering and text analysis technology. Because
now we don't necessarily have to write code to read text from all these
different sorts of servers that's locked into all these different
proprietary sorts of packets. CDF was one of the best things that
happened to CompassWare as far as allowing us to save a tremendous
amount of man hours, time investment, in effect, just as we used to do
with realtime news feeds, parsing all the different news feeds and the
non-standard formats. Now suddenly we don't have to do that with the Web
channel space. So we're pretty happy about what the Microsoft guys did.

>From the Floor: Hi. I'm from that delivery corporation, one of the
companies that supports the announcement that Microsoft made. I'm
wondering what happens with Netscape in this arena, and how will they
respond, and if they respond very aggressively, how does your
announcement benefit the industry? Because we'll still have more than
one set of standards.

David Waxman: Right. Well, obviously I can't speak for Netscape, so
they're going to need to do whatever they think is right. But we're very
committed working with the standards bodies, the WWWC and others. And we
think it's really important in this area, just as with HTML, that the
standard body be included in the creation of a mechanism for all content
companies to use, a set way of delivering information.

The example you give of a Netscape standard and then a WWWC standard is
something kind of scary. Because then that really works against the
ability for the industry to say, "Here's a consistent way," like HTML
today, right, HTML 3.2, to say, "This is my page and everyone should
display it equally." Whether you're Mosaic or IE or AOL or whoever.
Right? And I think that that's very important for all the members in the
industry to recognize and appreciate.

So I can't speak for them, but I would hope that they certainly work the
standards bodies as they should do.

Jeff Pulver: I mean, just to interject, in a totally different space, in
the space of telephony, H323 last summer, Microsoft and Intel announced
a standard for interoperability for the client software. And they went
out -- both Intel, Microsoft, the members of INTC -- and they promoted a
standard, H323. Well, guess who, a few months ago decided to also be
compliant -- Netscape.

So to the extent that Microsoft and others go to standards bodies and
put proposals out that universally add value to the industry, I think
you'll find historically, if there's a history here, that other vendors,
competitors specifically, also adopt, although they won't make a lot of
noise about it. And it's not really a me-too, it's just to the extent
that CDF adds value to the push industry, and Netscape has their own
agenda, to the extent that it adds value to what they want to do,
they'll support it, too. They may not say it up front, but it would be
silly for them not to.

It's not to say that they don't have their own agenda to push Java
applets or Java applications, but I fully believe that to the extent of
companies going out to the standards bodies and providing proposals that
make sense, there's going to be a consensus generated, and any one of
the players in the space benefit from joining in and agreeing to those
standards. So in the world of the Web, WWWC is the standards body. If
there's going to be a Netscape standard on top of that, that's news to
me. It's not to say that they won't innovate, but it's not going to be
where the consensus of the market's going to be.

Tal Saraf: And just to summarize with that. I think it's really
important that we recognize that if Netscape wants to make any changes
to them, they sit on the WWWC, they were given this proposal as any
other member of the WWWC, and they can make any proposal they wish.
Right? To enhance it, change it, revise it, whatever. In fact, with the
dynamic HTML specification, Netscape was one of the co-authors of the
specification, initially.

So this is an area where we're very happy to work with them. We go out
of our way to say, "This is the spec. We want to get peoples' opinions."
Our organization, the Developer Relations Group, actually had a design
review last week for both CDF and dynamic HTML. And Netscape was
certainly invited. They chose not to attend, but they were certainly
invited.

Peter Chislett: They were invited here tonight, too, to speak their
words as well. And someone just left the room and said it's too bad that
this turned out to be a Netscape-bashing event, so... (laughter) Sorry
about that.

But actually, I'm a big Marimba, a friend of Marimba, actually. Friends
of Marimba. And that's a technology that's not just in the Java space,
they're actually, right now, starting to take EXEs and distribute them
across intranets. They've been in beta for a couple of months now -- at
least a month. And I think IBM's heavy into that. So they're about to
open up those doors and allow their technology to be deployed.

I see a face on Jeff, and he's got a comment about this, I think.

Jeff Pulver: Nothing I can say publicly.

Peter Chislett: Okay. Well, I just think that what they have is
different from what everybody else has. Their tuner and their deployment
and their business plan, actually, is a lot different from a lot of
other people. And I think there's going to be a lot of space for them.
And of course for Microsoft there will be a lot of space, too.

Jeff Pulver: Okay, with that I think we'll transcend over to Peggy
Miles.

Peggy Miles: I wanted to thank Peter for pulling us all together because
it took a lot to get us all coordinated. And also for you for filtering
us and being our moderator today. My name is Peggy Miles and I have a
company called Intervox. And I am basically on the channel
implementation end for consulting and execution of things from live
events -- live broadcasts all the way to a couple of channels.

What I'd like to do is just briefly give you an overview of a couple of
channels that we're working on right now that are currently on the Net.
And then go through a process that we had to as a company, a lot of
others of us had to find resources -- my goodness, this is brand new --
what technology companies, what do we choose, who do we choose, where do
we find the content, what legal policies, what regulations do we have to
go to to actually push the data on the Web?

And first of all, I know a couple of people in here, I noticed there is
some webcasting search engines, some webcasters. We have Grit. We have
Jack Martin that's looking into live events. We also have book
publishers. What other kind of companies are in the room? Do we have
technology companies? Any kind of content creation people? Are you in
news, advertising, publications? Yes, sir.

>From the Floor: Nonsense. (laughter)

Peggy Miles: Nonsense. Good answer. We have a few more people in the
room. What else do you do? Cable? Cable modems? Cable industry? Cable
television? Web development and advertising on the Net through push
technologies. Anybody else that we haven't mentioned? Yes, sir?

>From the Floor: We're in advertising.

Peggy Miles: You're in advertising?

>From the Floor: Right.

Peggy Miles: Okay. Advertising agency? Okay.

Jeff Pulver: Can you define webcasting just for those who don't know
what it is?

Peggy Miles: Sure. Well, there's been a little of controversy over the
term "webcasting."

Jeff Pulver: Hence my reason for asking.

Peggy Miles: Good question. Probably in '95 I was lucky enough to be
involved with Xing Streamworks when they did the first major markets
broadcasting live in Seattle, and there really wasn't a term for this.
We really just talked about Internet broadcasting.

And then we changed the term over to kind of "webcasting" when we were
broadcasting or trying to push on the Web audio and video through
streaming technologies. And then when push technologies came through,
which is an enhanced form of data and text, and it can include audio and
video, it seemed to kind of all encompass everything in one thing.

I think Business Week on their front page mentioned webcasting as
basically a push technology, pushing it to your desktop. We generally
mention it as pushing things to your desktop, pulling things to your
desktop, interacting one-to-one with the consumer. And it can be audio,
video, and text. And it can be streaming, which is live broadcasting, as
well as audio-on-demand and video-on-demand, which is a very important
component of this -- is in this lifetime is how to give people things
they need.

The one thing that we all have in common is we lack one thing -- we lack
a lot of time. And as you've probably noticed, the society, if you're a
sociologist, and I work with human interfaces and graphical interfaces
and actually communicating, and what do these people want when they get
pushed to them, is we're all lacking time.

And what we're looking at when we're developing graphical interfaces and
graphical push environments is providing them with instant gratification
in the information they need when and where they want it. Which is using
tools, like the filtering capabilities. And a lot of the problems we've
been going through is developing this for companies is actually getting
it out of the content producer's location, because if it's a publisher
or broadcaster, they are not, or haven't been, Net savvy until the last
couple of years.

And that is a hard struggle in actually integrating this and talking to
the people that are actually on frontline in news departments, frontline
in publication departments and moving that to the Web. A couple of the
channels that I'm working on, and my company does consulting and
execution, is American Singles, it's a BackWeb push technology, and it's
also Cupid's Net.

So if you can visualize this, you go over and sign up for the channel
and you're interested in possibly a member of your own sex. And you sign
up the criteria that you would like which could be anything from
non-smoking, lives within five miles, certain socio-economic level,
race, whatever criteria you can ever imagine about somebody that you
would like to date.

And that is a push channel and basically what happens is this cute
little heart drops down on top. Yes, it can be obnoxious, it can be
intrusive, but on top of your desktop, whether or not you're working in
Word, Excel, or whatever it was, and this heart. See, if you click on
the heart, it says you've signed up for our channel. And we're now
sorting and we're finding the right matches for you.

So what I'm looking for is when I'm doing push technology, or finding
the channel and building it, is something that is going to be convenient
to a person when they want it and where they want it. And these people,
that might be really important to them to find somebody to live with for
the rest of their life. Just like it is to get the latest sports
information pushed to your desktop, are things that are very mission
critical pushed to your desktop that are filtered by companies like
this.

So the next thing they do is if you go to the site and you get this
other little heart, you click on it, and it actually drops down. I'd
love to show this, all of the men -- for me, I'm obviously a beta-tester
-- but, within five miles or four miles or ten miles of my home in
Washington, D.C. And I literally click on that person's criteria and it
brings me straight to American Singles' Web sites which allows me to
contact that individual instantly. But I already know that they match
and meet my criteria, so I don't have to go surfing through all these
Web pages. So I'm now combining interactive communications with database
marketing to push something that a consumer wants.

And that's our example, if you want to go to backweb.com or as.org. And
it's one of the channels for Cupid's Net, which is about 219 affiliated
singles channels. As.org stands for American Singles. The other -- and
this is confusing, too, because there are terms called "webcasting,"
"datacasting," for those of us who broadcast radio and television. Kind
of a pre -- earlier is the ability to datacast and move data through
certain portions of your television and certain other areas.

So I have a client, KCBS in Los Angeles, and if you go down driving on
some of the freeways in Los Angeles, there are four major billboards
that they have, that are 3-D billboards. They look like a big radio. And
in the radio, in the LED panel, is the actual name of the song that is
displaying in realtime on the radio station.

That was a form of datacasting that we actually used a signal to send
something to the billboards. Well, what we did is we took that one more
step further. We decided to take that same data -- by the way, getting
the data out of the radio stations and out of broadcasting areas,
because some of the equipment's analog and some of it's digital, and
there are human interfaces to get this stuff in realtime, is the most
difficult task that we have.

So what we had to do is we had to take that information, and it has to
be in sync with what's going on on the air. So if there's traffic on the
radio station, it has to be displayed on the billboard. Obviously, I
have limitations about how much I can actually display on a billboard
that's on the roads, because it's against the regulations that you can
flash a lot of text because it can cause accidents.

But what we did is we took that same datacasting and we pushed it to a
Web site through a Java applet. And we actually created this before Java
was on all the platforms, so we actually had a pearl version which was a
solid display. And what it did was it streamed the name of the song. And
it streamed the information that's going on concurrent with the
broadcast. And this application obviously can work with television and
radio. But if you think about it, what we are trying to do long-term is
provide a convenience to the user.

Now let's take the radio example. What do these people want in a local
community? They want local community information. They want weather and
sports. They may also, if it's an alternative rock station, want
information about that song because they don't know who the artist is.
So we are using the Web page as a resource or an information tool, so
they go, "Oh, I heard that advertiser's message. Where can I find out
more information?" For a radio broadcaster and for a radio audience, it
has to be what you remember, which is why radio is built on frequency
and impressions.

So what we're looking at is providing that information and all those
databases on the actual Web site for them in sync with a live broadcast
to enhance their experience of a live signal. And that's an example,
it's AeroFM.com.

And I'd like to give you a couple of information resources. It's kind of
fun because there is a number of you that I've only met for the first
time in realtime and basically a lot of you that I've met by e-mail.
There's a list service that you'll find on the third page, which is
webcasting. And there's about, now, 600 of us. It's in digest and it's
also in e-mail and includes the ITU Geneva, the National Association of
Broadcasters, the Canadian Association of Broadcasters, as members, the
Freedom Forum, it has product managers. Most of the product managers
from NetShow, Microsoft NetShow, which is what he mentioned earlier,
which is a tool which is Kodak-independent. In other words, for live
video and audio, or stored video and audio on the Web. Their particular
product at Microsoft can use any of the different technologies and some
of the technologies that are being used to push audio or stream audio
and video are things like RealAudio, RealVideo, Xing Technologies,
StreamWorks, and VDOnet.

So they have a tool -- and I don't want to be describing Microsoft's
product. But something to enhance the experience of people to try to get
further. They have some interesting things, when you have video, you can
actually click on areas that have hot spots. They have a neat example,
it's in a coffee shop in Seattle. And if you click on the back of the
sign it gives you the statistics and all the business information of
that coffee shop. So it's looking at a live video signal and actually
enhancing the text.

Anyway, we have about 600 people on the webcasting list. It's just a
public service from a lot of us that are trying to figure out things,
and we have a lot of lively debates, and you're welcome to join it. The
other thing that a bunch of us did back in September is we were totally
confused on all the technologies for our clients, we were totally
confused on what was going on. But we knew that audio and video -- just
like a lot of people have been predicting for '97 and '98, we work with
the IP Multicasting Association as well -- will be a force on the Net to
be reckoned with.

But with that came the problems with distributing audio and video across
international borders, including the things like the sovereignty of
nations and how to deal with that. We also knew that we that would be
restricted in places and we really didn't know who to talk to. So there
were a number of us that informally got together and formed an
association called the International Webcasting Association. And it
includes members like the Freedom Forum, CBS Radio, Paxon
Communications, Radio and Television, Xing Technologies' president, Mark
Cuban from AudioNet. It's a whole diverse group of people from content
providers, advertisers, to actually, we have the BMI and the ASCAP,
which are the music regulatory people. They're actually serving on a
committee together because they are in the same state of confusion,
because all of us are trying to figure out how to license our content
and distribute it to the Web.

There's information here, and I just wanted to thank you for giving me
the opportunity.

Jeff Pulver: Thank you. Any questions for Peggy? I have one. ASCAP, are
they the ones who are trying to get the Campfire Girls to pay royalties
for singing songs?

Peggy Miles: Yeah. There's obviously -- they are trying to support their
core people to make sure that they're paid the fees. And it just has to
depend on if this is a marketing process. A lot of radio stations are
objecting to that. And I think Grit over there is having a cyber
symposium on this coming up.

Jeff Pulver: Yeah, I remember getting some e-mail because I had a
RealAudio server, the folks from ASCAP wanted me to pay $1,000 to stay
legal. They didn't realize that the only content I was serving was my
own voice.

All right. Any questions from the floor? Oh, I'm sorry, on the side.

>From the Floor: You are part of webcasters.org.

Peggy Miles: Yes, one of the members.

>From the Floor: So if I go to webcasters.org would I be able to access
the list and basically gain information in terms of what the decisions
are for my product on a criteria basis?

Peggy Miles: That's a great question. We were organized in October.
We're having our first formal meeting at the National Association of
Broadcasters. And we have set up strategic committees that are run by
very high level people in the industry, from technology to advertising,
that will be producing that information and producing that on the Web
site. And sending it to people, their criteria, by e-mail on the latest
regulations from standards to legal issues to music licensing to
advertising. At least that's what we're trying to do, but we have to
bring enough people together from all these diverse industries to
actually talk to each other, and that's the hardest challenge. But yes.

Jeff Pulver: Peggy, could you just repeat her original question?

Peggy Miles: Her original question was if she went to the International
Webcasting Association -- webcasters.org -- would it be able to provide
her with certain information that might meet her needs as she is
developing her product? And that was the answer, that we are developing
the right committees with -- AT&T is on our technology committee as a
host, CBS is hosting our advertising, we have a content provider on
that.

>From the Floor: Do you still do work with Microsoft? Is Microsoft on the
committee? Is that going to affect -- I'm just curious -- the situation
at all?

Peggy Miles: We have requested Microsoft to join us, and Shannon Purdue
has contacted us by e-mail saying she's looking for the right department
and the right person to potentially be a member. So we are definitely
pursuing the browser companies and the technology companies and the
content companies and the legal and regulatory.

We also formed -- which is exciting -- IWA Europe, which is the person
that's running that is Flex Techs TV, which has Discovery Channel,
Bravo, and a lot of the huge Web sites. IWA Europe is already formed,
and we are working on right now IWA Canada, which we have a couple of
key members.

So there are very small groups right now, but we also have Digitmedia
Mims in Switzerland, we have Channel Africa, we have SABC. So we're
trying to get a diversity all over the globe because we feel that
webcasting will be not only large corporations, but individual
webcasters with content. And all of us are going to have questions. And
yes, we don't have all the answers. But we're looking for that group
that can at least help us develop the answers as we go.

Jeff Pulver: And just to summarize, we're not involved in that
association, to the best of my knowledge, at all. There are some program
managers who, I guess, are on the list server, but there are a couple
people on the WWWAC list server, too. Right? Thomas Riordan, who spoke
to WWWAC probably six months to a year ago, has been trying to respond
to the WWWAC alias, too.

So that's something that individual people will try to get on different
list servers depending on their interests, but that's really where it
ends. It's not trying to be involved with list server beyond that.

Peggy Miles: There's two things, there's a list serve and an
association. We have pursued Microsoft for the association. They have
expressed an interest to us specifically. We are just now -- we have not
formally announced them as a member and it's too early to do so.

>From the Floor: Just out of curiosity, could Netscape also be a part of
that?

Peggy Miles: Absolutely. We're just very brand new and we really haven't
been able to -- we've gotten a core group and we have just really not
even gone public with this except in limited quantities.

>From the Floor: I guess this would be an open question to all the panel.
Heidi, you've mentioned a couple of different applications in terms of
webcasting that you've seen. Or worked with really, in terms of recorded
music being broadcast over it, database [ ]. What have you folks seen
thus far as regards entertainment content created specifically for
webcasting push media?

Jeff Pulver: There is a lot of work being done in the adult
entertainment industry.

(laughter)

Jeff Pulver: I think they actually represent some of the best examples
of what this technology can do.

Peggy Miles: That's basically how VHS and VCRs got started, which was
adult industry. But yes there are individuals that have some dynamic
pieces of work, but they don't have the funding behind them or the
channel to distribute it. You don't hear about it. And it's just little
pockets of people. We had a request for somebody to be a member that was
a 14-year-old high school student for the webcasting association.

Peter Chislett: I just wanted to say again that this isn't a Microsoft
whole body up here -- we've invited Verity, CogniSoft, IFusion, a lot of
different companies. And they are all here for different reasons.

I have just one more announcement which...

(laughter)

It's like TV right? The first meeting of the WWWACs Webcasting SIG,
which is going to basically be, I believe, myself leading that one by
myself with Peggy Miles, will be our first person on the board. So you
can come and hear Peggy talk in a different room, smaller room, probably
in the city somewhere. Probably at the Lubin House, because that's where
we have reserved. 11 East 61st Street, Monday, April 21 between 6 and 8
p.m. And we'll probably get into more content-related issues.

So this was primarily for technical reasons, because it's a TechSIG. So
the WebcastingSIG is going to cover, I think, all facets of webcasting,
and we will talk more about content there. Again, it's Monday May 19 and
we'll be putting that up on the WWWAC.org Web site and also the TechSIG
Web site.

Jeff Pulver: Thank you. I might add that I also maintain at pulver.com a
pretty thorough resource of information for voice and video broadcasting
on the Net, as well as two-way communications. And I've, for the last
couple of years, been moderating a Voice On The Net list which doesn't
ignore webcasting but certainly doesn't take a look at all the issues.

About a year ago I also founded the Voice On The Net Coalition, a group
which includes Microsoft and Sprint and others. About a year ago, some
phone companies, not really understanding the technology, wanted to ban
Internet telephony, went to the FCC and asked them to regulate it. It
sounded kind of silly, but they did. So it took a couple of weeks and I
got a group of people together and we've been fighting them ever since.
Don't take regulatory issues lightly. To the extent that foreign
governments don't understand what this webcasting technology can do,
they will try to regulate it.

In a couple of weeks at a conference I'm producing in San Francisco, I
have an International Regulatory Panel coming out forth. Peter Harder,
counsel from Netscape, is chairing it. And we're looking at the chaos in
cacophony, and we have people from Europe and Asia looking at all the
issues dealing with the Net and seeing whether or not it's going to be
regulated like the television industry. Whether or not it's just going
to be a passing fad. To the extent that you have core technology out
there, don't take the regulatory aspects lightly. CDA and other things
happen. And it's not because you want them to, but it's because
legislators, people in Congress, don't understand where things go, and
the way that they deal with these things is by forcing regulations on
you rather than trying to understand the opportunities they represent.

>From the Floor: Tal, I looked through the CDF very quickly, the spec,
and there was the HTML part for the content, and there was a spec for --
which seemed to be sort of how often do you pull to see if the content's
been updated? Are there other pieces of that that we should know about?

Tal Saraf: Sure. I only brought one copy because I didn't know how many
people to expect, and it's on the Web. So I'd rather not kill extra
trees.

But there's a list of major elements, and I can kind of read through. So
the first piece is the channel, and you define the channel that way.
Next are various channel items, and you could think of that as
information that's available from a channel.

The next is sort of the user schedule, which is what you just
referenced, which is schedule updates. Then there's the schedule itself,
which could be a particular schedule that you would want to follow for a
channel. Then there's some logo information to represent the channel in
a minimized format.

Tracking information, if you wanted to provide tracking and obviously
tell the user what you're tracking. And then the category definition,
which defines the category and any potential children categories of the
channel. So those are sort of the high-level elements of CDF --

>From the Floor: Could you clarify the items in the categories parts?
Give an example of...

Jeff Pulver: Are there also any rating specs in there for content?

Tal Saraf: There isn't necessarily a ratings spec in here, although we
support the various rating formats and we encourage people to rate
channels. The channels that we would be including in IE by default would
be rated. That's a requirement that we are going to ask of people, but
beyond that you're sort of on your own.

If you want to take the item issue offline we can talk a little bit
about it. I can show you these packets. Kind of hard to do without some
visuals.

Jeff Pulver: Geo?

>From the floor - Geo: Yeah. I'm just wondering, maybe Jeff, you can
answer this, because you actually have a Web channel that I know of, and
what is involved to have that? What kind of organization, what kind of
preparation, and what kind of response do you get from having one Web
channel as opposed to being the other members on the panel?

Jeff Pulver: All right. I do confess. Back in October I set up a channel
on BackWeb called pulver.com. I publish something called the Pulver
Report which is my view on the world typically looking at the space of
broadcasting and telephony, but it certainly has interesting viewpoints
which are mine. I take no credit other than that.

Setting it up, though, is pretty complicated in the sense that it took
about $10,000 to buy a server. I got the software comp because I was
playing with it in beta. But it took about two weeks before I fully got
it up and running. But to the credit of the folks at BackWeb, the most
impressive part about the installation was the fact that this was
October and the documentation was dated in June.

So they actually had documentation, it was well thought out, and most of
the time when I look at beta software the documentation is to be
delivered in the future. So to the extent that there were people out
there and they actually had a package and it was just taking time to
disseminate, I was very positive about.

The time lags I had was just trying to interpret what it meant to the
different terminology that the folks at BackWeb had. Now when I update
my channel it could take 15 minutes, literally. I take a document and
HTML it. I then just put it through and I can push it through their
formatting sequencer.

Just to contrast that with Intermind, I also have a channel on
Intermind, and that took literally 15 minutes to set up totally. The
difference was that I didn't have to go out and buy a new Sun
workstation. I didn't have to get that configured. All I had to do was
download their client software and their publishing utility, follow what
really were simple directions. It did help that my content was already
HTMLed, and I just served it out.

And it's completely two different contrasts. As far as content goes, I
get comments on it. I think I am able to keep track of who the
subscribers are and it's a nice community of users. But I'm pretty
impressed with the way they can take control of my desktop. But I try
not to put out bitmaps and stuff that are terribly encompassing. But
overall it's been a positive experience, to the extent that they launch
more channels and more people find out about it, it's just interesting.

>From my own experience, though, as more and more channels come out there
-- part of the reason it attracted me to begin with was the fact that
six months ago push technology was new. Right? And the idea was, I was
differentiating myself by using that technology. Now as the space gets
very crowded and the 35 or so vendors out there, the idea of being
unique sort of goes away, and you sort of lose your identity. So I don't
know whether or not personally I'd be using that as my own marketing
tool in the future, because I try to look at things that are different,
and be there and then leave, and find the next best thing, rather than
let everyone else come up around me. So for my own personal purposes I
might use different technologies three months from now to publicize the
stuff I do. But I certainly see the industry growing very, very fast. I
hope that answered your question.

I'd like to thank everyone for coming tonight and appreciate your time,
effort, and energy and thank you to all the panelists.

Peter Chislett: Thanks for coming everybody.

----
adam@cs.caltech.edu

Just give it a little tap.
-- Happy Gilmore