lua-users home
lua-l archive

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]


On 6/7/2011 7:56 AM, steve donovan wrote:
> On Tue, Jun 7, 2011 at 2:19 PM, Patrick Mc(avery
> <spell_gooder_now@spellingbeewinnars.org> wrote:
>> It's true, it's just a hype meter, I actually trust this site
>> more: http://lang-index.sourceforge.net/
>
> It has a bias towards 'educational' languages, so under scripting
> you notice Logo ahead of Lua, in fact _Alice_ is ahead of Lua. (I had
> to look her up, and no she's probably not Alice ML)


Right, though if you look at the data, it's actually counting hits for both Alice 3D and Alice ML. Alice 3D also has a huge advantage on YouTube hits -- probably because no one says "Look at my cool Lua script results!" when they're posting videos of, say, their game that was written mostly in Lua.

The Yahoo results also look...wrong, somehow. Aha, I see how it's wrong. Yahoo was MAKING UP THE NUMBER. When you first do the search, sure enough, it says "75,000,000 results for Alice programming." But if you skip to page 10 of the results, it gives you an updated number that's presumably a bit more accurate..."393 results for Alice programming". So the initial estimate was only off by...FIVE orders of magnitude. Yahoo ends up with 556 results for +"Lua Programming", for comparison, if you skip enough pages forward.

What's even more fun is that if you do a Yahoo search for /Alice programming/ without the quotes and with no "+", it comes up with 9.9M results...I first thought maybe Yahoo was taking some combination of the frequency of the words "Alice" and "programming", and it may be, but the fact that the algorithm comes up with fewer hits for /both/ words on the page than the two words adjacent on the page just shows that the algorithm sucks.

Google kept insisting on "about 25k results" for +"Alice programming", by the way, as I jumped forward 10 pages of results at a time...until I hit page 41...when it suddenly became 404 results. Oops. +"Lua Programming" bottoms out at 419 results.

The moral of the story is: Don't trust the "about X results..." on Yahoo or Google to tell you about the popularity of a complex query. Both Yahoo and Google may be arbitrarily cropping results after you skip forward enough pages, too, since most people don't ever look past the first 2 pages of results. But at least Yahoo is certainly guessing at a number to print, considering the more specific query ended up with a higher number than the more general query.

Tim