Re: CGI, semicolons, and so on...

Rich Wiggins ([email protected])
Mon, 03 Jan 94 18:59:37 EST


>On Dec 30, 9:09pm, Charles Henrich wrote:
>>
>> I still say we go with the execute bit, 1000 times more flexible and simple
>> than any other method on the planet.
>
>What if I want to run a server on a DOS PC or a Macintosh? I think it is
>important to make sure that GCI is not OS specific. The continued growth
>of the Web is going to depend upon how easy it is to get new servers up
>and running. Not everyone is going to have access to a UNIX system for
>this.

I think this is a valid point -- the issue of how hard it is for new,
relatively naive users to set servers up is a factor that should
be part of all these discussions.

And for that reason it seems to me that the specification of a script
and its parameters should be explicitly, visibly different than
pointing to a file to be delivered. This discussion is a lot like
the question of whether a programming language should have implicit
typing or not. A lot of the arguments raised seem to tout opaqueness
as a virtue in its own right; although it doesn't seem necessary to
make it possible for users to tell that a script is running instead
of a file being served, I don't understand why it's important to hide
that fact from them.

It also seems to me that requiring all script-type documents to be
isolated in a "this is for scripts" subdirectory is unnecessarily
confining to the server author. Wouldn't it be better to be able
to store related flat HTML files and associated scripts in one
place?

It seems a suffix of .cgi may be a good compromise. Then it doesn't
matter whether the separator is a / or a ;.

/Rich Wiggins, CWIS Coordinator, Michigan State U