User talk:Sean Colombo/Archive/2007 Oct 6 - 2008 Jan 26
From LyricWiki
![]() Archive | |
|---|---|
New-Article Templates
The concept
Those templates are really brilliant. I'm surprised no one made "official" extensions for them (not to mention making them officially built in), so that you had to create your own.
Sadly, they are not documented in your other Wiki site, like the <lyrics>. You probably realized the Wiki system's greatest weakness is also its greatest advantage - its open system. The mass users out there are mostly newbies and they just can't handle an open system. Your templates counter that by putting (optional, but still) rules into the chaos. The autofill feature really tops it off.
To sum up, it simulates the order of a closed system (like, say, IMDb) with the mass users' appeal of an open system. So please share the documentation (if not even submit it as a feature request in Wikimedia) for a better wiki world out there. -Lwc 17:19, 20 November 2007 (EST)
Introducing the templates
How do you make an article using that extension?
- teknomunk (talk,E) 00:42, 6 April 2007 (EDT)
- Once it's up, just go to a new article. Until then, append &hax=sean_lwTemplates to the url of a new article.
- -Sean Colombo 00:44, 6 April 2007 (EDT)
- That's a good idea to have in there... but I'm not sure how we'd do that. Do you know of a way to get a Site Message from within the code?... similar to how {{int}} works in wikitext?
- -Sean Colombo 00:53, 6 April 2007 (EDT)
- Something like the pages on Special:Allmessages? It looks like MediaWiki uses the function wfMsg( $name ) to access system messages. For example: Skin.php:1086: wfMsg( 'aboutsite' ) ); is responsible for 'About LyricWiki' at the bottom of the page. Would this work? I think that these pages are protected by default, but I am not for certain.
- - teknomunk (talk,E) 01:03, 6 April 2007 (EDT)
- W00t, that would be perfect. :) We could just create some of those as the templates (and have something in there for replacements of $artistName, $songName, etc.
- Thanks,
- -Sean Colombo 01:39, 6 April 2007 (EDT)
Changes to Newarticletext templates
I would like to request some changes to those handy new templates.
- I would like to have an empty line before <lyrics> and after </lyrics> tag (and maybe those empty lines in between can be deleted? I would guess that many people dont remove them when filling out...)
- The templates create two empty lines at the end of each page. One is created by MediaWiki anyway, but at least the second can be deleted from the template, I think.
- With headings I would like to have a space character between the "="-signs and the heading's text (e.g. == text == instead of ==text==) to have a nicer looking source code.
BTW: Where are those templates anyway? I can't find them... --Speckmade 12:44, 12 April 2007 (EDT)
- The templates are part of an extension to the MediaWiki software and as such require server access to edit. This means that right now, only Sean can make changes to them. I previously made a suggestion to put the templates into the MediaWiki namespace like the items on Special:Allmessages, but the change has not been made yet.
- - teknomunk (talk,E) 12:50, 12 April 2007 (EDT)
Now what about the proposed changes? --Speckmade 21:06, 20 April 2007 (EDT)
Again - changes to Newarticletext templates
Sorry for the quasi-repost but you didn't seem to see it the first time and it's annoying me all the time and you are the only one who can change it since they "require server access to edit":
I would like to request some changes to those handy new templates.
- I would like to have an empty line before <lyrics> and after </lyrics> tag (and maybe those empty lines in between can be deleted? I would guess that many people dont remove them when filling out...)
- The templates create two empty lines at the end of each page. One is created by MediaWiki anyway, but at least the second can be deleted from the template, I think.
- With headings I would like to have a space character between the "="-signs and the heading's text (e.g. == text == instead of ==text==) to have a nicer looking source code.
- The inserted <lyrics> tags are not consistent with the <lyric> tag (withous the "s") the edit page recommends. Think we should decide (maybe prefer the shorter?).
Please respond. Greetz --Speckmade 17:10, 1 May 2007 (EDT)
- Hi,
- Oddly enough, I had space before the <lyrics> tags until someone asked me to take it out during testing... I definitely think this will have to be wiki-ed around with until we find a format most people like. I don't know how to add system messages to the Special:Allmessages page yet, but I've created the pages where you will be able to edit the templates. So that there aren't like 50 differently-formatted pages as we figure it out, I think we should work on it for a little bit until there is some kind of consensus on the template page before we start using the wiki-message live.
- Here are the links:
- MediaWiki:Newartist
- MediaWiki:Newalbum
- MediaWiki:Newsong
- Thanks,
- -Sean Colombo 02:59, 2 May 2007 (EDT)
- PS: Personally I prefer "lyrics" because it seems to be semantically/intuitively correct and I don't think the 2 characters per page are going to kill us on space. That's one of those things that will be good to iron out since now the "templates" (I shouldn't have used that word... it's already taken) have Talk pages :) Good idea on making these into editable wiki-messages!
- Well, I'm not able to edit those pages...
- I agree on the <lyrics>/<lyric> thing. (I wasn't really seaious with the space saving argument... thought there isn't any good reason for one or the other but you've found one.. :-) )--Speckmade 08:35, 2 May 2007 (EDT)
- To Sean:
- It looks like the templates are not being used yet. I did some looking on my own and found that you will need to use wfMsgForContentNoTrans("Newsong"); instead of wfMsg("Newsong"); so that the text is not changed at all.
- To add the message to Special:Allmessages, you will need to add $wgMessageCache->addMessage('Newsong', $message); to the extension or add the message in languages/Messages.php. I am not too sure about editing Messages.php, we should probably avoid changing the MediaWiki software too much.
- Hopefully this helps.
- - teknomunk (talk,E) 11:01, 2 May 2007 (EDT)
API
Searching for a song/album
Hey, I'm a HUGE fan of your site. I think its great you have created a lyric site with no annoying ads or surprising sounds that almost make you call the police=P. I noticed one thing that I think should be improved. When you search for a song, we'll use The Riddle by Five For Fighting for an example, if you search "The Riddle", you dont go directly to the lyric page. You get a sometimes long, confusing page that somewhere shows the link to that song. The same if you try searching the album name. Fixing this would be helpful if you know the name of the song, but not the artist. Can this be improved or fixed? -Richmond96
- Hi :) Although it's a challenge to modify MediaWiki to do this... it certainly can be done - and needs to be done. I think searching has been the weakest part of the site for a long time. Things have been really busy for us lately, but that is definitely becoming a higher and higher priority. I'm making this post glow green just to remind myself more visibly.
- -Sean Colombo 23:08, 31 May 2007 (EDT)
Thanks!! It would be great to see that happen.Thanks for responding so fast =)Oh, and i changed my account name.-Richmond313
Do you think that if you ever get around to fixing the whole search thing, that you could search for a user's name and have it come up to their page? I don't know, just a thought.-Richmond313
- This week, I added Google's SiteSearch to the left side-bar to replace the default MediaWiki search. So far I'm really pleased with the results. It doesn't do any jump-right-to-the-page stuff (because I can't modify Google's code :)), but it still seems to be a huge improvement.
- -Sean Colombo (talk|contribs) 11:30, 4 November 2007 (EST)
Link Tables and SOAP
I started working on getting Special:Lonelypages under control and came across a problem. It appears that when a song is submitted via the SOAP service, the link tables are not updated, causing several artist pages to be listed as orphans. The quick fix is to run maintenance/refreshLinks.php every once in a while, but the SOAP service should be fixed to update the tables automatically.
- teknomunk (talk,E,?) 12:57, 1 September 2007 (EDT)
- I just started refreshLinks.php running (looks like it might take a while)... I'll have to figure out a way for the soap to do it every time a new page is made. Hopefully there is just like a "refreshLinks(page_title)" type function that I can call after the page is made.
- Thanks for the info on this,
- -Sean Colombo 20:52, 1 September 2007 (EDT)
getArtist and capitalisation
Hi,
I have one minor note about <subj>. Api function getArtist returns at first sight correct results regardless of field artist capitalisation. But result only looks the same - links formed in api response will lead you to the correct pages only in 2 cases: exact capitalisation match or first letter mismatch. e.g.: searching for 'Muse' and 'muse', 'ABBA' and 'aBBA', 'David Bowie' and 'david Bowie' will produce good links, but bad for 'abba', 'Abba', 'david bowie' and so on. IMHO it's evident that service finds all required pages correctly regardless of capitalisation, but uses erroneouly capitalized search parameter for link formating instead of exact name of page was found.
regards, --Senvaikis 09:35, 12 October 2007 (EDT)
- Thanks, Sean, - now links are correct for all capitalization cases mentioned above (though it's still not clear for me, why search results can't be the same for 'MUSE' & 'muse' if they are the same for 'ABBA' & 'abba').
- One more note (I don't know, if it happened now or it was earlier too): in some cases response contains 'doubled' ArtistName parts. I suppose it happens when redirection occurs (but not for all redirects). I'm not sure, but seems to me that links are broken only in cases when redirected ArtistName contains more than one word. You can check this making search on 'Beatles', 'Rolling Stones' or 'Sensational Alex Harvey Band' (redirection to 'The ' +'ArtistName'). Redirection from 'Abba' to 'ABBA' does not produce any side-effects.
- Here's an example of such buggy link (using 'Rolling Stones' for search): The Rolling Stones:Aftermath (1966)
- --Senvaikis 07:03, 23 October 2007 (EDT)
- One more note: don't you think that getArtist response XML structure is slightly incorrect - wouldn't it be more consistent and parsable if nodes 'year', 'amazonlink' and 'songs' were childnodes of node 'album', not 'albums'?
- IMHO structure
- <albums><album><year /><amazonLink /><songs /></album></albums>
- looks more logical than current structure:
- <albums><album></album><year /><amazonLink /><songs /></albums>
- --Senvaikis 15:54, 27 October 2007 (EDT)
- Hi,
- As for the arbitrary capitalization, is that fixed now? There was a capitalization mix-up where I didn't apply the correct caps to the artist (only to the song), but I patched that up right around when you sent me the message. Please let me know if it's still broken.
- As for the XML format, I agree that wrapping the data inside of an album is more logical. I'll have to fix that in the next version (one problem we have right now is that I can't change the format on the fly since lots of apps are already implementing each function). If you'd like to use the properly nested format, let me know and I'll hack in a "version" parameter for now that will keep us tided over until I actually make a full new version (with better docs, versioning, etc).
- Thanks,
- -Sean Colombo (talk|contribs) 11:11, 4 November 2007 (EST)
- Thanks for response, Sean.
- As i've posted here already, I noticed your corrections to capitalization, and now it's ok.
- But you probably have skipped my second note (about broken links containing doubled artist names). Try this Api call for example.
- As for XML structure, it's not critical for me, but I think it shouldn't be very complicated task for you to add version parameter (or parameter nested=true/false with default false for old clients). Anyway, it dont seem to me as high priority job.
- As I've said, I treated all those notes as minor; but one question still remains extremely important for me - so dear Sean, please, be so kind, - answer at least shortly something about SOAP WS client in .NET (see here, here and here). If for some reasons this forum isn't a propper place for such discussion - may I ask you to drop me an e-mail?
- thanks,
- --Senvaikis 18:12, 4 November 2007 (EST)
- Thanks for response, Sean.
- Hi Senvaikis... I'll handle your points one at a time:
- Oops, I see it above that you said the caps were fixed already. Sorry.
- Thanks for the tip on the mispelled-artists bug that caused double-linking (and for the example link). That is fixed now too.
- Cool, XML will have to wait a little while then, I'm bogged down and can only get to fairly urgent things.
- In response to your .NET question about the SOAP, I haven't used .NET for this (and actually don't know of anyone who has used it for posting lyrics yet (several pull lyrics using .NET)... but I looked at the code and I noticed in one of your posts that you said you are passing an array of strings to represent album names. That is probably the problem right there. It may be a little confusing, but in the docs it shows that the onAlbums item is of type AlbumResultArray. So this should actually be an associative array with the artist, album, and year. I'm not familiar with C#, but as an example, in php this would look like: $onAlbums = array('artist' => 'Pink Floyd', 'album' => 'Dark Side Of The Moon', 'year'=> '1973');
- Hope that helps!
- -Sean Colombo (talk|contribs) 21:13, 4 November 2007 (EST)
- Thanks, Sean, - that's exactly I wanted to say! The problem is that WSDL declares onAlbums as string array! Look at WS method .postSong() how it's viewable in .NET object browser (e.g. - how it's declared in wsdl):
- Hi Senvaikis... I'll handle your points one at a time:
- (Notice, that parameter message also has wrong type of string). So, being .NET developer, I have no chance to declare postSong parameter onAlbums as type AlbumResultArray. IMHO it's evident that problem lays in SOAP WS wsdl file. If I'm wrong, I'd be glad to know where's my mistake :)
- regards,
- --Senvaikis 04:16, 5 November 2007 (EST)
- I've checked the WSDL file and it appears correct... here is the HTML version and the XML version of the WSDL. So I really can't fix it on my end. I haven't used .NET with WS, so I wouldn't know how to fix that, but one guy I know who does have experience using the LyricWiki WS in .NET is Cyber Sinh, who might be willing to help.
- Later,
- -Sean Colombo (talk|contribs) 09:49, 5 November 2007 (EST)
- thanks, --Senvaikis 12:31, 5 November 2007 (EST)
wrong behaving of ZSI
Hi,
I found you made redirect from Jaromír Nohavica:Zatímco Se Koupe?. This query is because error in python SOAP implementation or in web service. See LyricWiki_talk:SOAP#Unsing_webservice_from_python. This redirect don't solve this problem because I get back bad encoded lyric (as from all nonASCII songs). Is possible somewhere get complete communication for debugging? (I'll try to fix it because when I tried perl grabber, It was working and I can compare these two tasks and hope solve this problem).
Thanks
--PavelD 03:18, 27 October 2007 (EDT)
- Hi PavelD,
- The best way to get full debugging output is to use the really basic SOAP client and check the "debugging mode" box. It's really rough (ugly) still, as I initially wrote it just for myself.
- -Sean Colombo 14:20, 27 October 2007 (EDT)
Thanks. I have one further question. I analyze packets posted from ZSI and perl web service and only difference is in setting encoding for Soap XML request (may be in encoding of searched string too). (ZSI don't send any any declaration). Both of them send the same Content-Type (utf-8). The response in perl thread is in UTF-8 too but in ZSI is in ISO-8859-1. How your web service recognize the necessary encoding response? I'm asking because your web client post request in ISO-8859-1, but response is in UTF-8. This breaks my assumption that response encoding is affected by Soap XML encoding.
Thanks
--PavelD 03:56, 1 November 2007 (EDT)
- Hi,
- The response encoding is always UTF-8 since the data for the lyrics is in UTF-8 and can't be displayed properly in other encodings (since there are many foreign characters on the site). If you want to convert the response from UTF-8 to Latin1 in Perl, I'd recommend using the stringify functions of Unicode::String. It looks kind of like this (might want to search for the exact syntax, I'm not too sure on that last line):
use Unicode::String;
Unicode::String->stringify_as('latin1');
my $latin1String = Unicode::String::utf8($someStringThatWasUTF8);
- Hope that helps some.
- -Sean Colombo (talk|contribs) 09:36, 1 November 2007 (EDT)
It's interesting. Here is part of response I get:
<?xml version="1.0" encoding="ISO-8859-1"?> <SOAP-ENV:Envelope SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/" xmlns:si="http://soapinterop.org/xsd" xmlns:tns="urn:LyricWiki"> <SOAP-ENV:Body> <ns1:getSongResponse xmlns:ns1="urn:LyricWiki"> <return xsi:type="tns:LyricsResult"> <artist xsi:type="xsd:string">JaromÃr Nohavica <song xsi:type="xsd:string">ZatÃmco Se KoupeÅ¡ <lyrics xsi:type="xsd:string">Zatímco se koupeš ........ etc. ........ <url xsi:type="xsd:string">http://lyricwiki.org/Jarom%C3%ADr_Nohavica:Zat%C3%ADmco_Se_Koupe%C5%A1 </return> </ns1:getSongResponse> </SOAP-ENV:Body> </SOAP-ENV:Envelope>
I see the <lyrics> tag is in UTF-8 but the encoding at the first line is saying something else (ISO-88959-1) :-( If you are interested in, I'll send you whole dump.
--PavelD 10:43, 5 November 2007 (EST)
Server Stuff
mod_rewrite
The mod_rewrite problem has shown up again over on The Human League (Category:Label/A & M under the labels) I don't know what can be done. If you don't mind, could you e-mail me the settings so that I can do some testing on my machine. I could also use the code for the Batch Move, Lyric Extension, LyricWiki New-Page Templates and XML Formatter as I had to reinstall my OS and it wiped the code I had :/
- teknomunk (talk,E,歌) 23:10, 8 October 2007 (EDT)
- Emailed 8 files. -Sean Colombo 20:04, 9 October 2007 (EDT)
NFS
I noticed from in the IRC channel you were looking for help with NFS. I have sort of dealt with it,but only a VERY little bit, about the same amount as I have with SQUID :P I'll be in and out today, so just drop me a note if you need another brain to work at the problem (whatever it may be).
- teknomunk (talk,E,歌) 17:39, 25 October 2007 (EDT)
extention
Dear Sean, I see on Special:Version that the reason the web-site so coolly predicts what kind of template a new page should use is that you made an extension, LyricWiki New-Page Templates Extension (version 0.0.1). Is there any way that you'd be willing to share with another mediawiki developer? Pretty please?!?! (mrandmrsmurphy at gmail dot com) -- Åqúàŧĩkī - É - Ŧ 10:29, 13 November 2007 (EST)
- Sure :)
- I'm swamped at the moment and am digging myself out. When I get a sec, I'll make it available and I'll contact you. If I don't do this in a couple of days, feel free to bump me to get me working on it again (my talk page is long and I often lose track of things :[ ).
- See ya,
- -Sean Colombo (talk|contribs) 12:31, 13 November 2007 (EST)
- Great timing! Just last night, Teknomunk updated the template to make the defaults be 'system messages' so that admins can actually edit them right on the wiki! Pretty sweet upgrade.
- Anywho, I've released it on a new LyricWiki:Open Source page.
- Later,
- -Sean Colombo (talk|contribs) 10:43, 13 December 2007 (EST)
Weird
There seems to be a problem with the site. I keep getting red links for pages that exist or pages that are out of date from their current revision. Also, orphaned pages sometimes switches from one point in history to the current list. Do you know what's going on? Probably something to do with caching. I just thought I'd let you know. --WillMak050389 01:18, 15 November 2007 (EST)
- Yeah, something reeeeally weird is going on here. I restarted the memecache server, I thought that would have done it, but the site is still being weird. Diffs are showing up crazily... like when you edited this page, it showed Teknomunk as having edited it and it showed nothing new, just a bunch of removals. Recent Changes is coming up blank a bunch too.
- I have no idea what could cause this....
- -Sean Colombo (talk|contribs) 09:45, 15 November 2007 (EST)
- I've been noticing a lot of trippy little things too. Some things that I reciently edited were seemingly rolled back, but if you try to edit the page, the newer info is there. Also, page 'ghosts' seem to be popping up in the orphaned pages list, where they will be songs that I deleted or redirected, and do not actually exist, or have no edit history. Does this have something to do with the updates you have been doing serverside?
- King_Nee1114lyrics 15:14, 15 November 2007 (EST)
- Sean, wasn't November 6 when you created the database replica? If so, that would actually explain a lot. I don't really know the full server setup, but it sounds like it is switching between databases. Also, it looks like the databases are taking a beating right now; lots of 'Too many connections' messages.
- - teknomunk (talk,E,歌) 22:35, 15 November 2007 (EST)
- Good call. It was definitely crazy-broken replication. After talking with the WikiMedia guys for a while, it turns out that (although it's undocumented), the first server in your array of servers is used as your master. So we'd been writing to the weak server (which was meant to be the slave) and reading from it 99% of the time until yesterday when I set it to only 70% reads. The "master" was really just an out of date read-only server. I've switched everything around, so we really only should have lost about 30 minutes of data (right before I locked the site and made it read-only while I worked on the fixes... this was from about 3am-5am EST so very few people probably noticed the read-only lock).
- In the end, I think this means it will run faster now than before! (Since we were using the slow server basically as our only server).
- Thanks guys for reporting this so quickly and for helping me look into it!
- Let me know if you see anything else weird going on.
- -Sean Colombo (talk|contribs) 04:56, 16 November 2007 (EST)
Concerning Bj?rk
Something that I had noticed about the lonely pages is that certain artist appear, no matter how many other pages are linked to them, and also, in the case of Bjork and others, that the Other Songs page remains visible, despite being deleted. You click on the link, and there is nothing there.
And on a completely different note, there are some accounts that use an autogenerator for lyrics, and I was wondering if anybody knew where to get a program like that. (I mean sure, I could try to program something myself, but that has only ever ended in tears). The reason I ask, is that there are a few artists with fairly large discographies that are either not present at all, or very limited. By hand would not be the most effective.
If anybody has some info on either one of those, let me know
King_Nee1114lyrics 18:50, 30 October 2007 (EDT)
- Hiya,
- It looks like the Bjork issue has something to do with the way it's encoded. The link you're mentioning uses a %F6 to encode the "ö", but the normal way to encode it is %C3%B6 - I'm not really sure what the %F6 version means, or why it brings you to the correct page... hmm.
- As far as auto-generating pages, the SOAP API's postSong() method is probably a good place to start. I've spend the last several days cranking away with serious overhauls to UberBot so that he can submit another batch of lyrics soon (I'd made a personal goal to hit 500,000 good pages by Nov 1st (two days), and we're at like 465,000 now so I've got to get crackin'!). Hopefully this round of upgrades will include a way for postSong() to automatically add the listing for this song into the artist's page, but for now it will just sit as an orphan page for a while until the Janitor bot finds it.
- Thanks for your interest :) - If you have any sites to recommend as lyrics sources, let me know.
- -Sean Colombo 20:30, 30 October 2007 (EDT)
- Before the days of Lyricwiki, I used songmeanings.net and sing365.com. I would say 45,000 new songs in 2 days would be quite a feat, so good luck and let me know if you need anything.
- Going back to the other thing: is there a way to fix it and make them go away from the Lonely Pages?
- King_Nee1114lyrics 22:08, 30 October 2007 (EDT)
Uberbot Stuff
Überbot: Upping Open Systems Zeal
Hi Sean - How about having Überbot cold-shoulder iTune$, or more radically all walled gardens (vis' pages not accessible with open systems technology, e.g. requiring µCruft IE)? Thus, helping to starve evil empires out of existence. - Happy new year, Ian
Uberbot Code
Hey Sean, I was wondering if I could use the Uberbot code on my own mediawiki installation. I have a lot of content I could scrape off of the web, but I'm just learning RegEx and PHP. Is there anyway you could be persuaded to share/loan me the code? The stuff I'm interested in copying is all Public Domain. -- Åqúàŧĩkī - É - Ŧ 07:45, 18 October 2007 (EDT)
- Responded on Ŧalk page.
- -Sean Colombo 13:01, 18 October 2007 (EDT)
Not Categorized Yet
Self-released Albums
Hey, I noticed when I looked at the Quietdrive page, there were several albums that weren't acutally released by a record company, but they were self-released. So-- Anyway, my point is, if you are part of a band and have self released an album or EP, can you create an article for it? Quietdrive 17:48, 7 October 2007 (EDT)
- Absolutely! We encourage independent artists to add their music to LyricWiki. We feel that's one of the benefits of this site being a wiki: even the less-prominent bands can be included.
- Thanks for the question... I look forward to seeing your lyrics :)
- -Sean Colombo 18:31, 7 October 2007 (EDT)
Vandal
I didn't know where else to report this but I found a vandal and repaired his/her changes on one page: Umbrella:Umbrella... the ip was: 71.116.82.250 ... i corrected the changes and noticed it as it was on my watchlist since i had previously edited that page. Thanks. Congested 02:58, 9 October 2007 (EDT)
- Thanks, I warned that IP. It was their only edit, so I assume they were just testing it. I tend to assume that ppl aren't intentionally being annoying, they just doubt that they can actually edit the site when they first encounter a wiki, so they mess around with it to test it out.
- If you see that stuff around again, you can just put {{WarnUser}} on their talk page, and usually they'll get the idea.
- Thanks,
- -Sean Colombo 09:03, 9 October 2007 (EDT)
Edit to your user page
Hello. I edited your user page a bit to fix the floating in Firefox, and simplify the CSS a little. If you object, feel free to revert. —Pathoschild 04:11, 16 October 2007 (UTC)
- I dig it. Thanks,
- -Sean Colombo 12:03, 16 October 2007 (EDT)
SotD
Just wanted to let you know that Überbot did not update the Song of the Day for today. It usually will tell you if there was an error, but I didn't see one, so I thought I'd let you know. --WillMak050389 10:59, 20 October 2007 (EDT)
- Thanks! I probably wouldn't have noticed if you hadn't pointed it out. I poked him, and it's fixed now.
- -Sean Colombo 11:28, 20 October 2007 (EDT)
Thanks for the best Wiki ever!!!
I will transcribe anything you want!
Zana Dark 18:37, 24 November 2007 (EST)
Spam protection filter
I was trying to add a missing release to They Might Be Giants and when I tried to save I got the following message: "The page you wanted to save was blocked by the spam filter. This is probably caused by a link to an external site. If you think you entered normal text that we shouldn't block, please tell Sean or another admin.
The following text is what triggered our spam filter: *blocked out because it stops me saving this comment*" Now while the release I was adding didn't have that at all, however a song that was on the page previously did, yet this has never stopped me editing the page before. Well as the message says to tell you if it shouldn't be blocked, that's obviously what I'm asking. The Light6 23:16, 9 December 2007 (EST)
- Thanks for the detailed info! I've updated the spam filter and you should be able to freely edit TMBG without any more problems. Thanks for your contributions, and if you see anything messed up in the future, please let me know. :)
- Thanks,
- -Sean Colombo (talk|contribs) 23:39, 9 December 2007 (EST)
- Thanks for that. The Light6 21:01, 10 December 2007 (EST)
Wikify Tracklist
I had been using Lentando's wikify-tracklist-album-o-matic, but it has been unreliable at best lately. The frontpage is not broken, the server that does the calculation and output is. So looking into it further, I found it has happened before. Would we be able to host his tool on Lyricwiki itself? The reason I ask you and not do it myself is I (cough) don't know how. I had tried with a google page creator, but apparently google does not host PHP.
If it cannot be done, is there a similar tool somewhere else?
King_Nee1114lyrics 22:48, 21 December 2007 (EST)
- I just recently noticed it also. Sean created a nearly duplicate tool at wikify.php. I also wanted to mention to Sean one thing that I had noticed with it, seeing if they could be changed: Lentando's tool automatically capitalized all first letters of each song, this made it easier to convert all links to LyricWiki correctly-formatted links. --WillMak050389 02:16, 22 December 2007 (EST)
- Yeah, that tool should be more visible somewhere. I've been thinking there should be some kind of "Advanced Editor's Page" with links to that, the Special:Soapfailures, the iTunes Top 100 and other things of interest to the really hardcore users of the site. I'll make that sometime this morning.
- Good call on the ucwords(). Will do that this morning.
- -Sean Colombo (talk|contribs) 13:58, 22 December 2007 (EST)
- PS: Note that I'm taking 'morning' loosely here because it's already 1pm ;)
- I knew I remembered there was that wikify.php, I just could not remember where.
- Last night I had ironically stumbled across our local wikifyer, and used it to make some changes to Beck, but there are some things about Lentando's version that makes it a contender. I second the idea of adding capabilities to the local wikifier, such as automatic capitalization and other features.
- King_Nee1114lyrics 15:49, 22 December 2007 (EST)
- I've added the Help:Advanced page to make things like this a little easier to find in the future. I've also updated wikify to do the capitalizations automatically.
- Kingnee: you mentioned that there are some things about Lentando's version that you liked that might not be in our local version yet. What specific things would you like to see in ours? We can try to get those merged in pretty quickly.
- I tried to contact Lentando a while ago to offer to host his version after it went down, but I didn't hear back, so I don't have the original to compare ours to (and can't remember everything about it).
- Let me know,
- -Sean Colombo (talk|contribs) 16:12, 22 December 2007 (EST)
- Very cool, I am glad to see the advanced page up and running, with an easy way to get to the local wikifier. What I saw of Lentando's was first the ability to leave the alternate text (the title after the pipe) as unformatted (in lentando's, it is a checkbox option). Second is that his struck me as more clean. When you wikified, it would load a new page and show the output in a clean <pre></pre> box. Those are the existing features that I'd like to see integrated, but along the same lines maybe we could have also an option for {{Album Art}}.
- I cannot think up a longer wishlist right now, and I think that this is a very good start. It's a load off just knowing that we have something on our local servers, so it is as reliable as Lyricwiki itself.
- King_Nee1114lyrics 01:37, 23 December 2007 (EST)
- I found on Lentando's talk page that he left a link to the source for us. I've uploaded the original version in case you still like that one (but we should eventually roll all of its good features into the current script so that ppl can just go to one place regardless of where they got their input data). There is a link to the old version from the current version, and here is a direct link.
- Someone should give Lentando a barnstar or something for starting this whole idea, lol. I never sat down and figured out how to do those (if I did, I'd have given out at least a dozen by now, everyone on this site rocks pretty hard!)
- Later!
- -Sean Colombo (talk|contribs) 14:50, 8 January 2008 (EST)
Errors from SOTD update at 03:53, 24 December 2007 (EST)
Add More to Song Of The Day Queue
Automated warning: Sean, there are only 2 songs left in the queue. Please add more from the Song of the Day talk page.
-ÜberBot 03:53, 24 December 2007 (EST)
- Resolved -Sean Colombo (talk|contribs) 12:38, 25 December 2007 (EST)
Errors from SOTD update at 03:53, 25 December 2007 (EST)
Add More to Song Of The Day Queue
Automated warning: Sean, there are only 1 songs left in the queue. Please add more from the Song of the Day talk page.
-ÜberBot 03:53, 25 December 2007 (EST)
SOTD
Wanted to bring to your attention my recent nomination for Song Of The Day, since it's time-sensitive. I would imagine ÜberBot probably doesn't pay attention to such things. ;-)
— jF 05:46, 30 December 2007 (EST)
- Cool idea. I moved it over to the queue, so it should roll for tomorrow.
- Thanks,
- -Sean Colombo (talk|contribs) 13:45, 31 December 2007 (EST)
- I also bannered it on the front page, like last year (gee, is that already one whole year ago?). In two or three days, I'll just delete it or move it to the editors corner.
- As a side note: wasn't that album/song called "Super Trooper" instead of "Super Trouper"?
- Anyway, my best wishes to all members in general, and my fellow admins and Sean in particular. --Mischko
15:23, 31 December 2007 (EST)
- Thanks, guys. :-)
- Mischko, I was thinking the same, but noted when I was doing the nomination that it couldn't find the album the way I had it spelled. Went back and double-checked against the cover artwork, and obviously I've been spelling it wrong all these years. Apparently using British spellings instead of American was in vogue towards the end of the seventies. (Rumours, anyone?)
- Mind you, I now have this odd urge to spell it in all sorts of ways: "Sooper Trouper", "Souper Trooper", "Super Truper", even "Süper Troüper". God help me. ;-)
- Happy New Year to all!
- — jF 17:40, 31 December 2007 (EST)
Wikify tool
Just noticed over on the Community Portal that you are now handling the Wikify tool. I left a detailed comment on Lentando's talk page about that—mind taking a peek and seeing if that's something you might be willing to tinker with?
Thanks!
— jF 23:20, 7 January 2008 (EST)
- Shweet idea... I've added allmusic.com in there now. It doesn't do anything with track-lengths currently, but I did roll in that suggestion you had about counting tracks and giving a warning if that doesn't match the highest track-number we find as well as the change from brackets to parentheses.
- I just used one test-case (some Mary J. Blige album), so let me know if you find some bugs.
- Thanks,
- -Sean Colombo (talk|contribs) 14:41, 8 January 2008 (EST)
- So far, it's awesome. The only thing I've noticed that could be improved on is where AllMusic notes new versions and mixes (by using tags like "[#]" and "[live/#]" in the Title field). I didn't think to mention those in my post, since such tends to be rare. (I find them most frequently on reissues, where they use the single pound sign to indicate bonus tracks.) But you'll certainly not hear me gripe about making such tiny edits by hand—it's way less work than what I was doing before the new version of Wikify.
- Oh boy, I'm loving it. I threw Prince's "The Hits/B-Sides" at it (the most complex tracklisting I could think of on short notice), and excluding the aforementioned tags, it handled it beautifully. The track-counting trick for verifying a correct copy/paste works great as well. Gee, I feel like adding some albums. ;-)
- Mucho thanks!
- — jF 00:05, 9 January 2008 (EST)
- Glad you like it! I threw in the changes to remove the trailing [#] and [live/#].
- Thanks for adding discographies!
- -Sean Colombo (talk|contribs) 01:37, 9 January 2008 (EST)
- Oops, spoke too soon. ;-) Tried this album, and it didn't like it. Gave me only the album header line, with no tracks underneath. Also noticed that it apparently doesn't like apostrophes in the Album Name field—it will complete the album header line properly, but when the page reloads, it repopulates the Album Name field with "Neil" instead of "Neil's Heavy Concept Album".
- This album also breaks the track-counting (or maybe I should say field-counting). I notice that a paste appears to leave out any data before the first track number—maybe there's a rule involved that I've overlooked. Hmmm...
- — jF 01:13, 9 January 2008 (EST)
- Dang... I'll have to look at that.
- Thanks for the bug-report,
- -Sean Colombo (talk|contribs) 01:37, 9 January 2008 (EST)
- Coolness. It works for that album now, and it handles apostrophes better... it also will remember which mode you choose (Wikipedia, Allmusic, etc.).
- Let me know if it's broken for other pages too.
- Thanks,
- -Sean Colombo (talk|contribs) 20:13, 13 January 2008 (EST)
- Will do. Thanks again!
- — jF 22:27, 13 January 2008 (EST)
- Well... ;-) Just tried it again, and find that it dumps the first track entirely when AllMusic doesn't have anything in that very first field. Not a huge concern, since the field-counting kicks in nicely with a warning about it. Does the same thing with this album as well. Wonder if the tool could pull a further trick when it finds an incomplete copy/paste and try padding that first field with the appropriate null data? Maybe then the warning could change to something like, "Wikify detected a problem with the discography you entered. It has attempted to fix the error—please check the Wikified results carefully." Mind you, having to add a single song still beats the monstrous effort of editing by hand!
- Also noticed that AllMusic is rather inconsistent with how they flag tracks. In some cases, they use "[#]" for a bonus track, and in others (like that Fleetwood Mac album, they use "[*]". Like I said before, they aren't the best resource, but they do come in handy. Just wish they'd tighten things up a bit—it's annoying that they believe there's a FM song on that particular compilation titled "Make Me A Mash".
- On a side note, could you adjust the colors on the warning message to make it stand out a bit more? Something like yellow text on red background might be a little harder to overlook for us blind-as-a-bat folks. ;-)
- — jF 23:00, 13 January 2008 (EST)
Spam Filter Problem
you won't let me add Canibus lyric due to spam. The lyrics are long 200 bars. On the album Poet Laureate Infinity. How can I add them?
— Kevgauden 01:13, 10 January 2008 (EST)
- Those lyrics were really long so I didn't bother trying to figure out which word was in the spam filter... I just added them for that song (the filter isn't turned on for admins). I think this is the song you meant: Canibus:Poet Laureate Infinity Vocal 2.
- Thanks for pointing out this song,
- -Sean Colombo (talk|contribs) 21:52, 9 January 2008 (EST)
yeah it the correct song but the lyrics you've added are wrong. how can I put the correct lyrics on there.
- Dang, I messed 'em up! I dunno which part is being stopped by the spam filter w/o looking at the lyrics... can u send me a link to a page with them or email them to me?
- Sorry the spam filter is being lame,
- -Sean Colombo (talk|contribs) 22:01, 9 January 2008 (EST)
- Thanks! I put the update lyrics in there... my guess after skimming through the lyrics is that it has something to do with GlaxoSmithKline. The spam filter is mostly configured against adu1t sites and prescription drug spam.
- Someday in the future, maybe we can put a CAPTCHA in there if the spam filter is set off so you'd just have to type in some letters from a picture and it'd let you right through.
- Thanks again,
- -Sean Colombo (talk|contribs) 22:36, 9 January 2008 (EST)
SOTD nomination...
Not sure if it matters, since I'm sure it's going to get bumped to the Pending section anyway... But I noticed that a nomination got added to the top of the list, not at the bottom. Should such things be moved to the proper location?
FYI, the song in question is Tupac Shakur's "Changes", and should appear after the nomination for Lee Ann Womack's "I Hope You Dance".
— jF 23:42, 9 January 2008 (EST)
- Guess so. ;-) Thanks.
- — jF 23:49, 9 January 2008 (EST)
- I moved it down between Lee Ann and the next one & fixed the link to the song. Ppl probably just don't realize that we do them top-to-bottom sometimes... we mention it on the page but there really is a lot of text there :)
- In the future, you can just move them and maybe explain the move in the "summary". Users will probably catch on pretty quickly to how it works after they see a nomination moved once.
- Thanks,
- -Sean Colombo (talk|contribs) 23:50, 9 January 2008 (EST)
Anonymous spam
If you look at the anynomous edits list there is only spam from (probablly proxied) IP's. Appearantly we're on some spammers favorite list and I dare claim we have more work cleaning up these long lists of edits than there are good-willing anonymous users. I hoped it wouldn't come this far, but perhaps you should start considering blocking edits unless users are logged in. --Mischko
10:39, 14 January 2008 (EST)
- That would me doubleplusuncool to have to do that... but you're right that there's too much spam. On my lunch break I'll take one more stab at fixing this w/code.
- -Sean Colombo (talk|contribs) 11:26, 14 January 2008 (EST)
- In response to Mischko, I realize this is a problem right now, but unfortunately I've caught a few users doing the same spam (I'm assuming it is the same IP addresses) that are registered with a username. See this user's contrib. Obviously blocking IP edits would cut down on the spam, but that wouldn't eliminate it. Hopefully some sort of spam filter will eventually be able to remedy this, but if you remember, we used to have problem with spambots deleting all the content from pages appearing after an ampersand or percent symbol. See this edit. I fear if we fix this spam, they'll find another way to get us. That's why I think the CAPTCHA system for unregistered or new users would be a great idea. --WillMak050389 14:10, 14 January 2008 (EST)
- You probably also have noted that most of the spam consists of changes on one line, adding a random string of characters. Perhaps it would be possible to match edits by some filters (I think Wikipedia has a bot that automatically checks all edits after they are made), adding the pages with possible spam and their editors to a list which could be manually checked. If there could also be a link to block the user and rollback the change, it would reduce the task for an administrator to going through the list on one page, instead of going through the history logs. In any case, are there so much real edits by unregistered users that we might want to consider not blocking them? The problem, I think, with a captcha is that after entering it once, it won't stop spam -- or you'd have to require it more often (say, every n edits) which would probably work better but also hinder people who make a lot of edits (like administrators and users working on a certain artist). Actually, I'm not really sure there is even an effective way to deal with the problem, but I still believe (temporarily) suspending unregistered edits will work best; maybe after the bots find out they can't spam the site anymore they will leave. --Mischko
14:47, 14 January 2008 (EST)
- You probably also have noted that most of the spam consists of changes on one line, adding a random string of characters. Perhaps it would be possible to match edits by some filters (I think Wikipedia has a bot that automatically checks all edits after they are made), adding the pages with possible spam and their editors to a list which could be manually checked. If there could also be a link to block the user and rollback the change, it would reduce the task for an administrator to going through the list on one page, instead of going through the history logs. In any case, are there so much real edits by unregistered users that we might want to consider not blocking them? The problem, I think, with a captcha is that after entering it once, it won't stop spam -- or you'd have to require it more often (say, every n edits) which would probably work better but also hinder people who make a lot of edits (like administrators and users working on a certain artist). Actually, I'm not really sure there is even an effective way to deal with the problem, but I still believe (temporarily) suspending unregistered edits will work best; maybe after the bots find out they can't spam the site anymore they will leave. --Mischko
- I made a change that should drastically decrease the amount of spam that makes it through. I don't think this will stop the ATTEMPTS however. It took me a while to figure out what motivation a spammer would have to keep doing these vandalisms w/o any ads... clearly they're prodding the site to see if it's worth attacking, but wouldn't they realize very quickly that our site is too fast with reversions? After reading a bit about the state of spammers these days, my new guess is that these are just unintelligent bot-nets. These IPs might not even be proxies, they might just be infected computers (both are possible though).
- While I see your point Mischko about it being easiest to just make registration required, I still want to save that for a last resort. So here is my promise to you and the other editors: since I really want to make sure anonymous edits are allowed as long as possible, and I don't want to plague our community with reverting spam: I will roll out a new attempted fix for all new spam attacks that are reported to me within 24 hours of first noticing the report (I usually notice w/in a few minutes of the post), and will roll out fixes to more than half of these reports within 12 hours. If it turns out I can't deliver on this, then I'll probably revert to blocking anonymous edits.
- Imagine how sweet the internet would be if spammers wrote useful stuff instead? :-/
- -Sean Colombo (talk|contribs) 15:54, 14 January 2008 (EST)
- Something I've been meaning to ask, but definitely should given this topic. I caught a spam attempt in progress the other day, and trailed the user around reverting his edits almost as soon as he made them. I think he gave up after seeing several reversions.
- If we do notice an attack in progress, however, is there a way to see what admins are logged in so we can notify one of them immediately for any necessary action? I was going to let someone know, but had no way of knowing which admins were online. Didn't see any point in posting to a talk page, when they might not get it until some time after.
- If no such facility exists for seeing what admins are available, could we look into something to accomplish that?
- — jF 01:35, 15 January 2008 (EST)
- If you see a user spamming or vandalising pages, just drop a note on one of the admins' talk pages. For me, it doesn't matter which, because I watch them all via RSS, but I don't know about the others. Usually I notice vandal-bots when Category's start getting modified or an unusually large number of Japanese pages (we have a decent Japanese section, but we rarely get sustained editing).
- Hopefully I didn't rant too much and that was what you were looking for.
- - teknomunk (talk,E,歌) 03:42, 15 January 2008 (EST)

+ 

