Re: "Perl is the duct tape of the Internet" and other musings

Sundar Narasimhan (sundar@ascent.com)
Fri, 10 Jul 1998 07:56:35 -0400


Ron:

3 points;
1. I just got back from the "other side" of the world of high performance
distributed computing networks that exist today -- that views the Web as
a toy! Look up the performance requirements of networks like Saber's RES,
or SITA or ARINC's networks, or Teradata's banking stuff or phone
switching networks. [I often turn off OODB vendors when I ask them to
show me their reference site that's been running the longest without
any down time :) ]

2. Before I respond to the other posts, I thought I should just point out
one fallacy in Ron's and every subsequent post. It is simply not *wise*
to seek to go out and build "database-backed web sites" with a
"single" common set of tools (be they open source or not). The
requirements for building something like an e-store that just serves up
forms is quite different from interactive gaming (yes, some of these
have large databases for terrain, player profiles, saved games etc.).
So Ron, what kind of application are you thinking of building?

3. It is not necessary to specify behavior to do something useful,
even thought it might be nice. Look at sql-cli (the language vs. the
call level interfaces). Things like pl-sql and others exist, but
plenty work gets done without using behavioral stuff. The real
question to ask therefore is where is the equivalent of sql-cli in the
distributed-oo or xml-web world (before we get hung up on programming
stuff -- let's get the foundations done)?