I swear, one of these times, you're gonna wake up in a coma.

Cordelia ,'Showtime'


Buffistas Building a Better Board  

Do you have problems, concerns or recommendations about the technical side of the Phoenix? Air them here. Compliments also welcome.

To-do list


Noumenon - Dec 28, 2002 12:14:48 pm PST #2403 of 10000
No other candidate is asking the hard questions, like "Did geophysicists assassinate Jim Henson?" or "Why is there hydrogen in America's water supply?" --defective yeti

Adding a radio button for "Search thread #__" will probably happen between beta and 1.0, with 2.0 having "Search current thread" or "Search [thread names dropbox]."

Are you coding this, Noumenon?

Can I use Microsoft Foundation Classes? :) I only know some C++. I was just posting some lower expectations because I thought the coders might be thinking, "I would move Search out of beta, but I have to add the chronological sort and the snazzy search by thread function and it's a big job."


Jon B. - Dec 28, 2002 12:45:33 pm PST #2404 of 10000
A turkey in every toilet -- only in America!

I only asked because I'd never heard of that particular progression from Beta to 1.0 to 2.0. I thought maybe you knew something I didn't. ;)

Seriously, the difference between "Search thread #__" and "Search [thread names dropbox]" is very small. I would be against including the former at all. Thread number is very user un-friendly.


§ ita § - Dec 28, 2002 1:25:39 pm PST #2405 of 10000
Well not canonically, no, but this is transformative fiction.

Bingo! I just got CVS working.

Oy, they're lucky they're the only game in town.

Or maybe I should have read the docs first.

Anyhoo. I have to do a little more experimentation, and then we may be good to go.

Jon, I copied the current site code onto the test -- if you didn't grab the code beforehand, I have it.

Now, we have issues here. There's a repository, where the test code is controlled. This is good and necessary. But we need a testing methodology too.

Jon -- you up on CVS?

Who else do we have as volunteer coders?


Jon B. - Dec 28, 2002 1:28:43 pm PST #2406 of 10000
A turkey in every toilet -- only in America!

I know nothing about CVS, except that's where I buy shaving cream and razor blades.


§ ita § - Dec 28, 2002 1:38:40 pm PST #2407 of 10000
Well not canonically, no, but this is transformative fiction.

You might want to start reading. [link] is kinda overwhelming. One will need a client to access the server (http://wincvs.org), and here's a howto for the WinCVS client. Karl can give you login info.

I've already imported the files, and now I'm just trying to work through some more details.


Rob - Dec 28, 2002 1:40:37 pm PST #2408 of 10000

t volunteers

I can test drive CVS for you, too, if you want. I'm all about CVS.


Jon B. - Dec 28, 2002 1:57:51 pm PST #2409 of 10000
A turkey in every toilet -- only in America!

Should I be reading "the Cederqvist," then?

wincvs.org isn't resolving for me.


§ ita § - Dec 28, 2002 2:00:13 pm PST #2410 of 10000
Well not canonically, no, but this is transformative fiction.

Huh. Try [link]


Jon B. - Dec 28, 2002 2:01:46 pm PST #2411 of 10000
A turkey in every toilet -- only in America!

Nope. Not working either. I'll try again later. Maybe a problem with my nameserver.


John H - Dec 28, 2002 2:02:55 pm PST #2412 of 10000

I just tried to do a search for the term V!Giles and got hits for every reference to Giles. Should I put quotes around it or can I not search for the "!"?

I imagine that's because the indexing function splits posts up into "words", and one of the keys for the end of a word would normally be an exclamation or question mark.

I'm saying "words", not to do the "I" "am" "not" "ironic" thing, but because I found out very early when playing search engines that What Your Programming Language Thinks Is A Word May Vary, and indeed may vary between languages. Unfortunately we have invented our own ideolect where "Vamp!Giles" is a word, but MySQL doesn't think so.

Which reminds me, ita was talking about the obscure algorithm which determines which document comes to the top in a search.

I imagine it's actually something like the classic:

W(T,D)=tf(T,D)*log(DN/df(T))

where tf(T, D) is the term frequency of T in D. DN is the total number of documents df(T) is the sum of frequencies of T in every document considered or as it called the document frequency of T.

which translates to "the score for the term in the document, multiplied by it's uniqueness" -- it's not just that the term appears, but that it doesn't appear frequently anywhere else, that makes it score highly.