Re: Generalising inlined images (Kevin Altis)
X-Delivered: at request of secret on
Message-id: <>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 7 Oct 1993 11:11:13 -0800
To: (Rob McCool),
From: (Kevin Altis)
Subject: Re: Generalising inlined images
At 10:43 PM 10/6/93 -0500, Rob McCool wrote:
>I put the INC stuff in because it seemed like a useful feature. It was not
>one of my more popular decisions around here. I didn't make it an additional
>type since I was not aware that the INC tag would be of any value to the
>I'm starting to wonder what exactly this feature is used for that is not
>better satisfied another way, since:
>1. Parsing every HTML file adds overhead to the server
>2. The only applications I have seen are things like including the date
>(which should be in the protocol) and Charles Henrich's weather server.
>So my main question to those who want this support in the server is, what do
>you use it for?

So far, I've done the following with <INC>, many of which might be better
as actual server commands. However, each one took me a few minutes to
provide cross-platform information access to a large base of users. I'll
never be able to do that if I have to provide a script or C program under
/htbin for every thing I think of and more importantly your basic user
won't be able to do any of this stuff without the server administrator
getting involved - for those of you that have never done System
Administration that should be read as "sorry, user, you can't have that."
1. <INC SRV "|date">
        Everyone does this right?
2. <INC SRV "|stock | /usr/bin/tail -9">
        This runs a program which goes out and gets the hourly stock quote
for Intel via a UDP "connection" then the results are piped to the "tail"
command because I only wanted the last 9 lines of output.
3. <INC SRV "telnet domain port | /usr/bin/tail -14">
        This gets the last week of stock activity via a telnet connection,
much like providing information via a finger lookup.
4. Using <INC...> on text and HTML documents so that I have controlling
documents with a <HEAD> section, then all the stuff I <INC> have only
<BODY> stuff, so they can be included in a lot of different documents
without conflicting with anything else. This is called document reuse, much
like code reuse. Obvious elements for reuse are parts of your profile or
signature, standard disclaimers, quotes, block quotes, etc.
5. A source code overview system does an <INC> which grabs part of some
source code files to show changed elements in the code (via greps,
compares...). The HTML document with the <INC> tags doesn't change,
everybody still accesses the same HTML document, but the body of the HTML
document potentially changes each time somebody loads it.

Well, that's a start. I haven't had enough time to get creative, geez I
only got the capability a week or so ago. I forsee using <INC> mostly for
including information dynamically in a document, like having <INC> provide
the default field entry for a FORM or using <INC> along with ISMAP to
provide a dynamic map front end where the click on an image causes a
command to be run at the server which generates output, which is then
incorporated into a document that is static, except for the <INC> elements.
Maybe I have to show you that one. Suffice to say, I find <INC> more useful
than many of the other tags.