Bounce, bounce, bounce...

Robert S. Thau (rst@ai.mit.edu)
Thu, 09 Oct 1997 20:11:27 -0700


I've been trying to get email to you, but it keeps bouncing. So, I
thought I'd append a "where are you" to a Fork post... and it
bounced. Sorry I missed you this weekend at Redbones... but I got in
late that evening, and dinner was probably over by the time I got the
message. The Fork post, btw, follows, for what it's worth:

Subject: Michael Malone says Bill has won. Has he?

A pointer to an interesting little thing from Upside, for those who
haven't seen it already:

http://www.upside.com/texis/features/companies?UID=97110110013

The thesis of the article, by Michael Malone, is that Silicon Valley
has become hollow at the core because Microsoft and Intel are, by
now, in a position to pick winners and losers, and to absorb or crush
anybody who gets big enough to deserve their attention, at will.
In this, he's even more pessimistic than I am. Still, at the very
least, the guy gets off some good lines --- among which is his take
on the "summit" last summer (whose main significance he reads,
correctly, as being less in anything that Bill told the CEOs of the
top firms in the country, the elite of his own industry, and the Vice
President of the friggin' United States, than in that he could summon
them all at will). Or this:

The most common question asked in the hallways of
Silicon Valley, Manhattan and even Seattle is: What
does Bill Gates really want? The answer, which we all
know but refuse to admit, is that he wants everything.

And then there's the wonderful one-liner, "Cassandra was right".

That said, life has gotten more interesting for Microsoft over the
past few days. I'm referring less to The Lawsuit, than to the
avalanche of complaints I'm seeing on the Java newsfroups about
IE4.0's poor Java support. There are AWT bugs, of course --- even
though they claim to have JDK 1.1 event model support, their browser
fails to run one of the JDK 1.1 sample applets at all, and has
problems with some of the others.

The problems go beyond the libraries, though: for instance, this
method --- a skeleton of something larger which had mysteriously begun
to fail ---

public int bugRoutine(int i) {
for (;i < 10;i--) return(1);
return(-1);
}

apparently returns -1 in IE4.0, but in everything else out there
(including IE 4.0 betas), it returns 1 (correctly, in accord with the
Java specs); this seems to be due to a novel JIT bug.

Does anyone out there have a feeling as to whether these complaints
are representative of the actual community reaction out there? If
they are (which is, I admit, a very big if), then MS may have a bit
of a problem here.

Like I said, I don't think Microsoft has much to fear from the
courts, at least not soon --- the only thing that can happen that
would *really* hurt them would be an injunction a few months down the
line against using the Java name, and even that's a long shot. (I
suspect that an injunction regarding Sun's trademark would do them a
lot more damage than anything relating to shipment of derivatives of
Sun's code --- there are enough clean-room VMs out there that MS
could buy one up quick if they had to, or they could buy the company
that wrote it, for that matter).

However, if a lot of CIOs refuse to install IE 4.0 because of poor
Java support, and they all tell MS that "you support Java well when
Sun says you do", *that's* a problem for MS.

(Nor is this the only reason to worry about installing IE 4.0, BTW;
there are also reports of IE 4.0 installs completely zorching
peoples' machines. But that's a whole different kettle of scrod).

But hey, I'm in Boston, so what the hell do I know?

rst

PS: I'm not defending every aspect of Sun's behavior here by any
means --- ISO was perfectly right to reject Sun itself as an "open
standards body". But I know which poison I'll choose if I have to.

PPS: Rohit --- if you get this, sorry, my mail's been bouncing.
What's a good current address?