2003-05-11 14:44:43

Dear Google, Please be careful.

Dear Google,

Please be careful.

The rumor mill says you are thinking about removing weblogs from your main index.  Some of us are concerned about this.  We like you.  We visit you every day.  But we want to remind you that your success has come because you consistently deliver what we want.  So just in case it isn't obvious, here's what we want:

As much as possible,
we want you to
show us the web without affecting it, and
index the web without organizing it.

It might help ease our concerns if you would tell us why you want to separate weblogs from other websites.  For now, I can only speculate:

• Perhaps you believe the quality of weblog content is a lower so it should be searched separately?
• Perhaps your indexing farm wasn't designed to handle tens of thousands of sites which are updated on a quasi-daily basis?
• Perhaps you are simply trying to provide a better search for weblogs, like you did with catalogs, newsgroups, and newspapers?

I suppose I wouldn't mind having an index for weblogs only.  In fact, I might appreciate the ability to search with results constrained only to weblogs.  I already do this using rssSearch.  But I am more concerned about the notion of removing weblogs from your main index.  When I search the whole web, why would I want to constrain my searches to exclude sites that you happen to call weblogs?

Regardless of your reasoning, I want to know how you will decide what is a weblog and what is not.  It's not a trivial question.  Do you have a perfect, unambiguous, black-and-white definition of a weblog?  I don't.

How will you describe all the other sites that continue to remain in your main index?  This isn't USENET with Deja.com, where the lines actually were "black and white".  Weblogs are websites, and the boundaries around this particular subset are not always clear.  How can a website manager predict which label you will apply to his/her site?

Here is my ten-part definition of a weblog:

1. A weblog is a website.
2. The items in a weblog are usually presented in chronological order.
3. A weblog is usually written by one individual in the first person.
4. A weblog often posts links to other sites of interest to the author.
5. A weblog usually has an RSS feed.
6. A weblog is usually updated regularly, daily or several times per week.
7. Some weblogs have places for readers to enter comments.
8. Some weblogs have a blogroll.
9. A weblog usually lets the personality of its author show through.
10. A weblog usually contains content produced by an amateur writer, not a professional.

Note that criterion #1 may be the only thing that all weblogs have in common.  We all agree that weblogs are websites.  After that, things turn gray very quickly.  All of the other nine criteria include a fuzzy word like "usually".  I can't imagine using these ten things to draw meaningful boundaries.  Exceptions are easy to find.

• For starters, I assert that my own site is a weblog, although I don't update it every day, I don't support comments, and I don't have a blogroll.
• Is Scripting News a weblog?  Obviously.  What about DaveNet?
• Is Six Log a weblog?  Surely, yes, but it's not written by one individual.
• We all agree Dan Gilmour's site is a weblog, right?  But he's not an amateur writer.  Would you make an exception for him?
• Is Joel On Software a weblog?  I think so, but the quality of content there is unusually high, including the complete text of a published book and enough stuff to produce two or three more.  If you're going to filter websites by quality of content, surely you're going to make an exception for Joel, right?  Who else has content which is just good enough for the main index?
• Is Cafe Au Lait a weblog?  It looks like one to me, but it has no RSS feed.
• Is Slashdot a weblog?  I don't think so, but some people do.  It is listed on weblogs.com.
• To be fair, I'll confess I can't find a weblog which fails criterion #2 above.  It seems that all weblogs present their items in chronological order.  Perhaps criteria 1 and 2 are the real definition of a weblog.  So, that means MSNBC is a weblog, right? 
• Is MSDN a weblog because it has RSS feeds?
• Is the home page of the Mono project a weblog?  Its items are presented in chronological order, and it supports an RSS feed.
• If a corporate site were to change its format, showing daily news about the company's products, in chronological order, with an RSS feed, would it become a weblog?

Thanks for recognizing that weblogs are interesting.  In many ways, the web is simply a new medium for old things, but weblogs are actually new.  Nothing like them has ever existed before.  Weblogs are one of the few areas where the web is a voice rather than an echo. 

Please be careful.