Personal Website of Yakov Shafranovich

« Sinking to the Lowest Level in Journalism         A Standalone Web Application »

When Non-Standards Collide: Mozilla’s Prefetch, the NoFollow Tag and Google’s Web Accelerator

(This post was part of a separate “Standards Blog” which has been merged into my main blog)

For quite some time, Mozilla-based browsers such as FireFox, Netscape and Galeon had a feature called “pre-fetching”. What this feature does is that it enabled browsers to “pre-fetch” any webpage link marked with a tag “rel=’prefetch’” to make the browsing experience seem faster. Google even added a feature to their search results to take advantage of that.

Also, back in January, Google in conjunction with many other search engines and weblog software vendors announced a new feature to combat comment spam – the “no follow” tag. Marking a link with “rel=’nofollow’” tells search engines to ignore the link in results, thus killing the very purpose for comment spam.

Now last week comes a new Google Labs creation – the Google Web Accelerator. This little tool prefetches and caches pages to make the browsing experience faster. However, this time not everything went as smoothly as before – as pointed out on the Signals Vs. Noise Blog causes many problems. However, the only way to turn it off is the same way as turning off prefetching – by looking for a special HTTP header. Note that regular prefetching as described above is set via an HTML link but can only be turned off via an HTTP header.

There are several interesting questions. First of all, prefetching as defined by the Mozilla FAQ refers to the specific act of prefetching web pages marked with the “prefetch” link. However, Google’s Web Accelerator is using the same header EVEN for links that are not marked that way. Second, among the suggested solutions for some of the problems caused by Web Accelerator was to use the “nofollow” tag to indicate to Google that specific links should not be fetched. BUT there is no way to do that since the “nofollow” tag is used for search engine rankings. Of course we can defined a new “noprefetch” tag type but that would only help if Google follows it.

Unfortunatly, what is missing from the discussion so far is one glaring fact – none of the “tags”, “headers” or “behaviors” described here are standards of any kind. Instead, all of them are proprietary extensions by vendors, must alike some of the non-standard behavior often used by Microsoft. However, this time since this are the darlings of the Internet – Google and Mozilla, there is no outcry. Unfortunatly, there are not standards – rather their are “non-standards”. A lot of the issues that are being described by people with these tools are due to the way different people do things on the Internet. Standards are ways to get around that by having an common agreement.

So what’s the solution? It is time for all the parties involved to sit down and formally standardize all of this stuff. All of the companies involved are members of the W3C and should be able to submit this for formal standardization while working out the thorny issues. Otherwise, these will simply devolve into a “Wild West” game of “non standards”.

Tags: google, mozilla, w3c —

Permalink | Trackback URL | This post has