Of absolute URLs and gzipped files

[Matt]

After reading the article entitled "Some URLs are better than others" mentioned in this thread, I went through and removed the absolute linking. Individual files (e.g. page.html) became index files within a directory (e.g. page/index.html) and links had the index.html portion removed. There are a few problems w/ this:

1. Dreamweaver no longer tracks when you move a file or change the directory
2. Dreamweaver doesn't display site previews properly
3. Advice here suggests that absolute linking is better (at least for Google)

I also went through and compressed every html file w/ gzip. Now many of the various tools for checking web sites (key word analyzers, validators, etc.) are only detecting the gzipped content, but aren't properly decoding it.

So, I can deal with annoyances #1 & #2. #3... does it matter if my links are broken in Google's cache? Does this affect page rank?

As for the compression problem: Rather than relying on the browser to properly report whether it supports .gz files, refer to a list of browsers known to be compatible. Whereas relying on a browser to properly report whether it supports .gz via a one line addtion to .htaccess described here, my intuition tells me that referring to a "browser whitelist" would be more involved. Any recommendations/ references on how to do this?

These are considerations for others thinking about doing any of the above. Maybe someone already has a solution to the gzip problem... if not, I'll post back my solution (assuming I create one) if anyone is interested.

[jeep]

On the absolute v. relative linking debate - I don't think there are definitive answers to that. I've flip-flopped several times on the issue. A lot of it depends on how much portability you want for your code.

As for gzipping... I do it using jpcache.php which asks the browser if they support zipped content. If you aren't using server-side scripting to ask this kind of question I'm not sure how Google and others will handle it.... Not sure how much help that is!


Jason

[Jeff]

3. Advice here suggests that absolute linking is better (at least for Google)

The only note regarding absolute linking I could find was:

Use ABSOLUTE url references in all your pages. Google caches those pages, and when it serves up the page out of its cache, the only way some of those images are going to get loaded is if you have absolute references

which does not refer to the /subdir/index.html or /subdir/ debate. All it seems to be saying is that you should absolute link your images if you want them to load property from google's cached version. But this does not seem to be true either, as relatively linked images on my pages seem to load fine from google's cached version of the pages.

[Jeff]

I was surprised to see .shtml grouped with .asp/.php/etc on the no-no list for being google friendly.

I would also be interested in an easy summary for how best to serve gzip pages in addition to html pages. I started to read the apache docs, but then Terra mentioned in another thread that current versions of IE still may have a problem with this, so I put it on hold for a while. But it still might be on my to-do list as I would love to have my static html files load as fast as my vBulletin which uses gzip. And a nice way to save some bandwidth too, possibly with a nightly cron job to create the updated gzip versions of html pages which have been changed.

Edit: oops - didn't see your other thread. Following those links now.

[Grendel]

What exactly was the searchengineworld.com article's reason for saying that dynamic page extensions (*.shtml, *.asp, *.php, *.jsp...) were not Google-friendly? That's news to me... my website's pages are dynamic and they work just fine in Google. If Google couldn't handle dynamic pages properly, it would be one badly broken search engine.

Frankly, I would take that article's advice with a grain of salt.
"3. Join a link pyrimd program to build links. Yes I still recomend this. "
"7. Put up walls between (like Excite) major topic areas (try not to link back to your base page from a subsite)..."
I mean, c'mon.

The most amusing part is that the article links to
another article on the same site about Page Rank, calling it "highly informative and good reading for the search engine addict." The "informative article" turns out to be a long circa-1998 screed about how Google's newfangled Page Rank system will never work.

Some of the advice in the original Alistapart article is top-notch, however. For example, the point about hiding your pages in directories -- I have done this for over two years and it is great. You can convert your site over from html to php with 'nary a hiccup. (I think of this idea as vaguely like object-oriented programming: you hide your implementation so that if you make changes, it doesn't screw up your external users.)

However, the point of THAT is simply to make the maintenance and administration of your site easier. If you are using Dreamweaver to help your administration, and Dreamweaver doesn't like that scheme, then the point is moot! So you should ignore it.