| View previous topic :: View next topic |
| Author |
Message |
FastTurtle
Guru
Joined: 03 Sep 2002
Posts: 500
Location: Flakey Shake & Bake Caliornia, USA
|
 Posted: Fri May 27, 2005 2:42 pm Post subject: |
 |
|
Why not look use the Google Search function as TLDP.org does and let google worry about what terms are going to be ignored.
This would definately improve the quaility of search results and reduce duplicate threads as we will be able to find the information in the forum.
_________________
AsRock B550 Phantom Gaming 4
128GB 3200 Mhz memory
1TB NVME as the boot disk
4x 4TB Sata - 2x 2TB Sata SSD - 4x 450GB SaS - 3x 900GB SaS - 72GB SaS for Gentoo system disk
LSI 9300-16i in HBA mode for all spinning disks
Radeon 6800 (Non XT) for GPU |
|
| Back to top |
|
 |
tecknojunky
Veteran
Joined: 19 Oct 2002
Posts: 1937
Location: Montréal
|
| Posted: Sat May 28, 2005 12:23 pm Post subject: |
|
|
| amne wrote: | | Given M. Sur wrote: |
Anyways, I just wanted to voice my objections. Thanks for reading. |
We are of course aware that the stopwords list isn't the perfect soltution and has some limitations. However we think that the positive effects outweigh the negative ones. |
Not to abuse of your time, but could you elaborate this? Because I have a hard time beleiving this.
Yes, before their was maybe lots of words indexed and yeilded irrelevancies in search, but I don't see how it compares to be a better solution than to eliminate relevancies as it does with stopwords. 
At least, please tell me some other solution is in the work.
_________________
(7 of 9) Installing star-trek/species-8.4.7.2::talax. |
|
| Back to top |
|
|
amne
Bodhisattva
Joined: 17 Nov 2002
Posts: 6378
Location: Graz / EU
|
| Posted: Sat May 28, 2005 1:20 pm Post subject: |
|
|
I'm not much into the database stuff myself, there's a post by klieber here. The stopwords would increase the size of the database a lot, which is bad.
_________________
Dinosaur week! (Ok, this thread is so last week) |
|
| Back to top |
|
|
tomk
Bodhisattva
Joined: 23 Sep 2003
Posts: 7221
Location: Sat in front of my computer
|
| Posted: Sat May 28, 2005 6:05 pm Post subject: |
|
|
| tecknojunky wrote: | | amne wrote: | | Given M. Sur wrote: |
Anyways, I just wanted to voice my objections. Thanks for reading. |
We are of course aware that the stopwords list isn't the perfect soltution and has some limitations. However we think that the positive effects outweigh the negative ones. |
Not to abuse of your time, but could you elaborate this? Because I have a hard time beleiving this.
Yes, before their was maybe lots of words indexed and yeilded irrelevancies in search, but I don't see how it compares to be a better solution than to eliminate relevancies as it does with stopwords.
At least, please tell me some other solution is in the work. |
At the time the extra stopwords were added the entire forums were pretty much unusable due to the very poor performance. This was one of many measures used to increase the performance and usability of the forums. The words were chosen as they were the most frequently occurring words found in posts and their relevance was therefore greatly reduced.
Bare in mind that matching up which words are found in which posts is the greatest database bottleneck in the whole forums, the stopwords are just one of things that we've used to reduce this bottleneck.
_________________
Search | Read | Answer | Report | Strip |
|
| Back to top |
|
|
FastTurtle
Guru
Joined: 03 Sep 2002
Posts: 500
Location: Flakey Shake & Bake Caliornia, USA
|
| Posted: Sat May 28, 2005 8:29 pm Post subject: |
|
|
Please don't add "solved" to the stop words as it's one that I use to find problems that have already been solved. Sometimes they have the solution to a problem I'm having or there's enough info to remind me how to fix the problem my self.
_________________
AsRock B550 Phantom Gaming 4
128GB 3200 Mhz memory
1TB NVME as the boot disk
4x 4TB Sata - 2x 2TB Sata SSD - 4x 450GB SaS - 3x 900GB SaS - 72GB SaS for Gentoo system disk
LSI 9300-16i in HBA mode for all spinning disks
Radeon 6800 (Non XT) for GPU |
|
| Back to top |
|
|
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20585
|
| Posted: Sat May 28, 2005 9:26 pm Post subject: |
|
|
| FastTurtle wrote: | | Please don't add "solved" to the stop words as it's one that I use to find problems that have already been solved. |
It'll eventually add itself when it becomes a very common search term... it'll cease to be valuable.
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
benny1967
Apprentice
Joined: 25 Apr 2004
Posts: 224
|
| Posted: Sun May 29, 2005 9:11 am Post subject: |
|
|
| pjp wrote: | | FastTurtle wrote: | | Please don't add "solved" to the stop words as it's one that I use to find problems that have already been solved. |
It'll eventually add itself when it becomes a very common search term... it'll cease to be valuable. |
I think this shows a misconception about how searches work, both technically and from the usability point of view. A search term doesn not become useless only because its common. (What you probably mean by 'useless=common' search terms are words like "the", "a", "but" etc.) A word may be a valuable search term because it's common.
What would happen if google put all the words from http://www.google.com/press/zeitgeist.html on the stopwords-list?
You are, of course, right in saying that it doesn't make a lot of sense to search for "error" or "gentoo" (assuming they are common terms) alone. it might make a huge difference, though, wether I serach for "error xchat connect", "howto xchat connect", "xchat connect solved" or "xchat ebuild error compile". It's the combination that matters, not the words.
I do, of course, realize why stop words need to be used as long as the hardware can't cope with the increasing number of posts and searches. But: Please don't start making this a "it's too common, so it's of no use" thing. Otherwise, chances are that the stop words will remain even when (at christmas) the hardware will change 
BTW: i noticed there are some german words in the stop list; therefore, i strongly believe that "man" got there because it's a german word for 'one' (in the sense of 'one would rather not use stop words...'), not because of the man-command. it might be useful to remove it so that searches in the english language forums can find references to the man-pages... |
|
| Back to top |
|
|
Given M. Sur
l33t
Joined: 03 Feb 2004
Posts: 648
Location: No such file or directory
|
| Posted: Sun May 29, 2005 9:43 am Post subject: |
|
|
| benny1967 wrote: | | it doesn't make a lot of sense to search for "error" or "gentoo" (assuming they are common terms) alone. it might make a huge difference, though, wether I serach for "error xchat connect", "howto xchat connect", "xchat connect solved" or "xchat ebuild error compile". It's the combination that matters, not the words. |
Exactly. That is why the stopwords have seriously degraded the search function's functionality.
Oh well. I'll just use google for now on.
_________________
What is the best [insert-type-of-program-here]? |
|
| Back to top |
|
|
tecknojunky
Veteran
Joined: 19 Oct 2002
Posts: 1937
Location: Montréal
|
| Posted: Mon May 30, 2005 8:20 pm Post subject: Try to find this thrue search... |
|
|
Just another answer that cannot be found (exatly this one) because... well, you know why. Can you guess? Have you at least tried? The forums problem got from bad to worse. If you can type a big question, someone might find it... probably, or maybe not. oups, bug.
_________________
(7 of 9) Installing star-trek/species-8.4.7.2::talax. |
|
| Back to top |
|
|
legine
Guru
Joined: 27 May 2004
Posts: 555
Location: Germany
|
| Posted: Tue May 31, 2005 1:55 pm Post subject: |
|
|
I am not sure if the archievments of stopwords have solved a thing.
Maybe it is a solution for now but I guess in one or 2 years we see a similar problem with the Foren. And then it gets harder to do anything. Maybe a problem comes up and everybody writes textes with evolution, Or we get a Open source X replacement named Windows for Linux
The problem is that a lot of Programs do have common used words. To keep them away from the stoplist there has to be a "vote" or at least a offical place where you can object to the words.
No admin can keep all the nice Apps in mind to think about, if you know what I mean.
So we would need a own offical forum page where Words can be viewed and rated by each member in order to get a clear view on the words which are needed and which are not.
Maybe if its just a Softwareproblem we could splitt the Forum in smaller parts. I.e. Generate a seperate Instance for every langauge and make a seperate one for the most frequented english one (like Off The Wall).
We would need then a Metasearchengine which is abel to search every Forum the user likes to search. But that one should be around somewhere too 
This of course is just Ideas what we could do in order to solve the Problem. It is not perfect in my opinion but a way to think about
I do not think that oracle will help since Oracle is complicated in its own ways
Another Forumsoftwere might delay the Problem. Even the String search is just something for the moment. We have 2.4 Million posts and I bet the number does not shrink. Even if some Artikles get deleted due Time I think this forum get more Artikles in then are aged out due time (if that happens anyways.)
Cheers
Legine
_________________
quote from Spaceballs:
Dark Helmet:[...] we were told to comb the desert, so we're combing it! [puts down bullhorn] Find anything yet?!
Soldier: Nothing yet, sir. |
|
| Back to top |
|
|
ian!
Bodhisattva
Joined: 25 Feb 2003
Posts: 3829
Location: Essen, Germany
|
| Posted: Tue May 31, 2005 2:58 pm Post subject: |
|
|
All we need is a better search engine. And yes: We're doing some research on this.
We know about the downsides and know about the problems.
_________________
"To have a successful open source project, you need to be at least somewhat successful at getting along with people." -- Daniel Robbins |
|
| Back to top |
|
|
legine
Guru
Joined: 27 May 2004
Posts: 555
Location: Germany
|
| Posted: Tue May 31, 2005 4:19 pm Post subject: |
|
|
What is the status on the reserch? Are there any results by now? Who is doing the research?
I am just a bit curious. Sorry for asking 
_________________
quote from Spaceballs:
Dark Helmet:[...] we were told to comb the desert, so we're combing it! [puts down bullhorn] Find anything yet?!
Soldier: Nothing yet, sir. |
|
| Back to top |
|
|
ian!
Bodhisattva
Joined: 25 Feb 2003
Posts: 3829
Location: Essen, Germany
|
| Posted: Wed Jun 01, 2005 6:13 am Post subject: |
|
|
| legine wrote: | | What is the status on the reserch? |
Reading a lot of docs and articles, looking at alternative search functions/engines.
| legine wrote: | | Are there any results by now? |
No.
| legine wrote: | | Who is doing the research? |
Me.
_________________
"To have a successful open source project, you need to be at least somewhat successful at getting along with people." -- Daniel Robbins |
|
| Back to top |
|
|
Given M. Sur
l33t
Joined: 03 Feb 2004
Posts: 648
Location: No such file or directory
|
| Posted: Wed Jun 01, 2005 8:26 am Post subject: |
|
|
| ian! wrote: | | legine wrote: | | Who is doing the research? |
Me. |
You rock ian!
Of course, I appreciate all the admins' work, not just ian!'s
_________________
What is the best [insert-type-of-program-here]? |
|
| Back to top |
|
|
mc_barron
Apprentice
Joined: 28 Aug 2003
Posts: 230
Location: Chicago, IL
|
| Posted: Thu Jun 02, 2005 3:58 pm Post subject: |
|
|
What about having some sort of labelling for each post? Or perhaps a forum divided up by portage package (like bugzilla, but specific to the ebuilds in portage)? So if I can't get <something> to compile, I can go directly to a set of messages directly pertaining to <something> instead of searching for the word "something"?
Or something... |
|
| Back to top |
|
|
fabs_uk
n00b
Joined: 01 Jun 2004
Posts: 15
Location: university, the joys of
|
| Posted: Fri Jun 03, 2005 7:29 pm Post subject: |
|
|
hmm, flickr style tags for posts? admittedly there's no backwards compatabilty, and it puts a *whole load* more stress on the servers, but it'd be way cool
<ahem>
right, i'll go and try do something productive for once (unlike this post!) |
|
| Back to top |
|
|
blubbi
Guru
Joined: 27 Apr 2003
Posts: 564
Location: Halle (Saale), Germany
|
| Posted: Mon Jun 06, 2005 8:06 pm Post subject: |
|
|
I can fully understand the problem of the growing DB and that there has to be done somthing. But let me add my notes about this stopwords.
The search for a problem has grown much harder. In 50% of my searches I get "no posts found" or just useless results.
Most of the time I search for a error message like "perhaps you want to do kbd_mode -u". This resulted in a good posts list because most people post the entire error message. like
| Code: | * Loading key mappings...
loadkeys: warning: this map uses Unicode symbols
(perhaps you want to do `kbd_mode -u'?) |
So if the onyl word that counts is "kbd_mode" you'll be lost in 1000 of results. But entering the phrase "perhaps you want to do kbd_mode -u" now results in "No topics or posts met your search criteria".
But indeed, there is a thread using EXACT that phrase (https://forums.gentoo.org/viewtopic-t-308102-highlight-kbdmode.html)
Since the stopword list was introduced, searching on the forum is no longer effective. It's just annoying.
I don't know exactly if the performance would be different when using a commercial solution such as VBB. I remember the poll and discussion to switch to an other type of board, and YES, my vote was to swich to a commercial solution.
They STOPWORD-WAY is deffinetly the wrong way. Switching to a larger server is no sollution, but may be switching the forum-software would be a good start.
The way it's now it's no way.
Sorry for not contributing a sollution but criticising.
But I know you/we gonna solve this. This community just rocks!
regards blubbi
_________________
-->Please add [solved] to the initial post's subject line if you feel your problem is resolved.
-->Help answer the unanswered
http://olausson.de |
|
| Back to top |
|
|
tecknojunky
Veteran
Joined: 19 Oct 2002
Posts: 1937
Location: Montréal
|
| Posted: Tue Jul 05, 2005 4:52 pm Post subject: |
|
|
How does one search about php throught Google? With site:forums.gentoo.org my custom php problem, the php term will bring up the scripts of phpBB (ie: viewtopic.php).
_________________
(7 of 9) Installing star-trek/species-8.4.7.2::talax. |
|
| Back to top |
|
|
Given M. Sur
l33t
Joined: 03 Feb 2004
Posts: 648
Location: No such file or directory
|
| Posted: Tue Jul 05, 2005 10:11 pm Post subject: |
|
|
| tecknojunky wrote: | | How does one search about php throught Google? With site:forums.gentoo.org my custom php problem, the php term will bring up the scripts of phpBB (ie: viewtopic.php). |
| Code: | | allintext: my custom php problem site:forums.gentoo.org |
Unfortunately, that will not weed out the links that people have added to the page. But, it will at least weed out the hits that are just in the URL.
_________________
What is the best [insert-type-of-program-here]? |
|
| Back to top |
|
|
tecknojunky
Veteran
Joined: 19 Oct 2002
Posts: 1937
Location: Montréal
|
| Posted: Wed Jul 06, 2005 1:24 am Post subject: |
|
|
| Given M. Sur wrote: | | Unfortunately, that will not weed out the links that people have added to the page. But, it will at least weed out the hits that are just in the URL. |
Woohoo. Thanks
I'm getting more and more relevant results from these forums now. It's a bit long to write, and I guess one could bookmark Google+site:forums.gento.org.
_________________
(7 of 9) Installing star-trek/species-8.4.7.2::talax. |
|
| Back to top |
|
|
mdshort
Apprentice
Joined: 06 Dec 2004
Posts: 157
|
| Posted: Mon Jul 18, 2005 5:27 am Post subject: |
|
|
This is VERY frustrating when searching for specific results, you people need to delete most commonly searched words from that stoplist, because I frequently become very frustrated and impatient with the search engine when I need to find something specific to a category, when something as simply defining as "server" or "mail" can't be searched for. Google has this feature too but atleast they do it right.
_________________
"With every rise, there is a fall." |
|
| Back to top |
|
|
pkluss
n00b
Joined: 10 Nov 2003
Posts: 27
Location: Illinois
|
| Posted: Mon Jul 25, 2005 4:13 am Post subject: |
|
|
I know that people have made this perfectly clear, but the search function is incredibly frustrating. I love the Gentoo forum, the information within is priceless, but I'll be honest, I've given up on the search function. 90% (no exaggeration) of the time it takes a thoughtful query of mine, chews it up, and spits out a mere skeleton of the initial submission. More often than not, no useful results come back either.
The main problem for me seems to be that any search I put together that is slightly related to the next gets pared down to the exact same search in the end, even though the subtle differences should have been enough to offer discriminating results.
I won't claim to know the solution here, and I'm all for clever solutions, but this one just doesn't work. Please focus efforts on revamping this, I think it would be WIDELY appreciated throughout the Gentoo community. For the first few days I literally thought that it was broken.
In closing, I love everything Gentoo has become so far, it never ceases to amaze me. I'm sure that this will be solved as well, but let's do it sooner than later.
Thanks. |
|
| Back to top |
|
|
drescherjm
Advocate
Joined: 05 Jun 2004
Posts: 2790
Location: Pittsburgh, PA, USA
|
| Posted: Sun Jul 31, 2005 4:34 am Post subject: |
|
|
| Quote: | | I know that people have made this perfectly clear, but the search function is incredibly frustrating. I love the Gentoo forum, the information within is priceless, but I'll be honest, I've given up on the search function. 90% (no exaggeration) of the time it takes a thoughtful query of mine, chews it up, and spits out a mere skeleton of the initial submission. More often than not, no useful results come back either. |
I could not have said it better myself. I came back to this thread to find the instructions on how to use google to search because the search function is as it is today almost completely useless in my mind. I just did two queries and both threw out most of my keywords so that I got hundreds of results when I expected only a few. 
_________________
John
My gentoo overlay
Instructons for overlay |
|
| Back to top |
|
|
legine
Guru
Joined: 27 May 2004
Posts: 555
Location: Germany
|
| Posted: Sat Aug 06, 2005 2:38 pm Post subject: |
|
|
Have you checked http://getvanilla.com/
I did not looked deeper myself, but the mainpage sounds promising.
I do not want to interfear with your work, I just saw this and thought: "hey maybe it is a good shot".
I apologize if this pointing is not wanted. And tell me what information you would need to accept them I see if I can organize them.
_________________
quote from Spaceballs:
Dark Helmet:[...] we were told to comb the desert, so we're combing it! [puts down bullhorn] Find anything yet?!
Soldier: Nothing yet, sir. |
|
| Back to top |
|
|
Corona688
Veteran
Joined: 10 Jan 2004
Posts: 1204
|
| Posted: Tue Sep 06, 2005 11:03 pm Post subject: |
|
|
There's at least one upside to the restrictive search keywords. "STFU n00b" style responses are quite rare because there's no way anyone could reasonably be expected to find useful information here unaided.
_________________
Petition for Better 64-bit ATI Drivers - Sign Here
http://www.petitiononline.com/atipet/petition.html |
|
| Back to top |
|
|
|
Gentoo Forums Feedback
