Thursday, January 25, 2007

Another Search Oddity (this time Google Groups)

Earlier this month, I ran across some apparently unusual search results using Microsoft Live Search. I still think they were peculiar but I wasn't able to pin anything down conclusively. See the posts here and here.

Now I have something odd from Google Groups (i.e. their Usenet archive) in the form of content that used to be there but now appears to be missing. I understand that Google sometimes removes content for copyright and other reasons, but I can't imagine that's the case here. It's just an old post from the comp.os.ms-windows.apps.utilities.win95 newsgroup discussing an old DOS-based program of mine, Directory Freedom.

I was going through some old paper files and I ran across a hardcopy of an old (1996) posting that heaped (rather overly effusive) praise on my program. The hardcopy was a printout of a Google Groups search result. I can't tell the exact date, but there's a 2001 Google copyright notice on the bottom so I'd assume it's from 2001 or not much later. (A scan of the hardcopy is here.) Anyway, just for kicks, I thought I'd pull up the link and save a copy of the page. And I couldn't find it by searching Google Groups.

I'm not sure what to make of this. I first tried searching using some unique terms from the post and didn't come up with anything. Then I reverted to using the same search string as I had originally ("gordon haff") and going through all the results by hand. Still nothing. As far as I can tell, the post is just gone (or at least removed from search results).

Have I missed something obvious? Or is there some other explanation for this result?

[UPDATE- Mystery (mostly) solved thanks to Brandon who tracked down a search string that did work. (See comments.) My name is mis-spelled in the post in question. So why did it show up in my previous search? I can't say for sure but perhaps Google was just returning "looser" search results five years ago when there were a lot fewer sites and posts than there are today.]

3 comments:

Anonymous said...

Here's a link to the post, not sure how you had issues finding it, I do notice that searching by the authors email address apparently doesn't work.

http://groups.google.com/group/comp.os.ms-windows.apps.utilities.win95/browse_thread/thread/f7d68cc877552ded/7489477de424af9c?lnk=st&q=group%3Acomp.os.ms-windows.apps.utilities.win95+directory+freedom&rnum=1#7489477de424af9c

Anonymous said...

If that link isn't completely visible, its the first result by searching for:

group:comp.os.ms-windows.apps.utilities.win95 directory freedom

Gordon Haff said...

Thanks Brandon,
I do see the post based on your search. You've hit the nail on the head with "I do notice that searching by the authors email address apparently doesn't work." After I couldn't find the post using my name, I plugged in the post author's name and that didn't produce a result.

That still leaves the question of why it appeared in the prior search and not this one. But closer examination reveals that my name is mis-spelled in this post. So, perhaps, when I searched previously Google was being more forgiving of deviations from the literal search result--which makes sense given the growth of the Web during the period.