Text Only Version

[zmroberts]

I have been asked to make a "text-only" or otherwise dial-up-modem friendly version of a large already established corporate web site. This site is huge and uses various technologies (asp, java/jsp, static pages). Making an entire copy of the site without the images is not really a good option. Is there any type of software solution or does anyone have any ideas?

Thanks

[moltar]

One way I can think of it - you can proxy all the requests and filter it with Perl.

[zmroberts]

this seems to be similar to something that the BBC web site has done with "BETSIE" (http://betsie.sourceforge.net/) ... one question.. if the site I am working on resideds on multiple servers... some IIS and some UNIX/Websphere... would the perl still work? What exactly do you mean proxy?

Thanks

[moltar]

You can configure Apache to act as a proxy for other requests. This is usually used for big websites with multiple servers.

User requests a page from Server A. Server A fetches the request from Server B and pass it on back to the user. The process is transparent to the user. Server A may want to pick a server from a server farm (load balancing). In your case, server A can fetch the request from content server, pass it through a Perl script/module which strips all unnecessary code. I think this can be accomplished with mod_rewrite, mod_perl and PerlResponseHandler.

I guess in your case writing the stripping script would be the hard part. I am sure if you look around you can find either Perl modules that can do that, or some command line utility or something of sorts. I can imaging this problem came up many times.

[limbo]

Another option, if the site is built with CSS, is to use style switcher that provides a text only or low graphic version. This is really easy to do with Javascript. But of course that depends on how the site is templated... tables?

[zmroberts]

yep ... lots of tables and lots of content created by different people so all different styles. So using the CSS would require "touching" every page to add certain div tags or whatever around all the images, right?

[moltar]

So using the CSS would require "touching" every page to add certain div tags or whatever around all the images, right?

It most likely be the other way. You'd have to edit each page and remove certain tags. Probably most of them. It's a tedious process.

[zmroberts]

yeah... that's what I figured... and that really wouldn't work for me. I really need to figure out a way to dynamically create a seperate web site based on bandwidth... I like the idea of the perl program but would also be curious if anyone else had any alternate approaches?

[moltar]

You might not even need a separate server. If you are running everything on a dedicated machine, you can configure the text version to run on another port. And then filter everything through that port.