Hacklog: Blogamundo — poking holes in the language barrier since approximately 1 month from now

b
l
o
g
a
m
u
n
d
o

Global Voices / Blogamundo IRC Meeting Log

Written by Patrick Hall, 2 years, 8 months ago. Tags: No Tags.

Conversation with #globalvoices

(14:57:22) The topic for #globalvoices is:Conversations and arguments around the world, since October 26, 2004…
(14:57:46) patfm :describing blogamundo
(14:58:18) patfm :actually we just put up a blog and i wrote a post about, well, here,hehe: http://blogamundo.net/dev/
(14:58:58) ethanz :cool - let me get through that and I’ll be in a better position for ourchat in a few moments…
(14:59:24) patfm :k great
(14:59:29) patfm :sorry it’s kind of long
(14:59:59) ethanz :love it. it’s a great introduction to the conversation
(15:00:19) ethanz :I’ve been working a little bit with some friends who are working on acomprehensive localization framework for open source software
(15:00:23) ethanz :some of the same issues apply
(15:00:28) ethanz :distributed lexicons, etc.
(15:00:32) patfm :right
(15:00:39) patfm :i’ve found that there’s tons of work on localizatoin
(15:00:43) RMacK :hi sorry…
(15:00:48) patfm :but not so much about straight ahead translation of content
(15:00:58) patfm :heya RMacK
(15:01:23) ethanz :right - the localizaztion stuff is mostly a catherding problem - thetranslation stuff involves more work getting people psyched about theissues…
(15:01:28) RMacK :hi.
(15:01:35) ethanz :hey RMacK…
(15:01:38) RMacK :hey
(15:01:39) BorisAnthony [n=bopuc@pdpc/supporter/active/Bopuc]entered the room.
(15:01:50) patfm :right… our theory is that the key to getting more translation done isto lower the barriers just a bit
(15:01:52) patfm :heya boris
(15:01:59) BorisAnthony :heyas
(15:02:00) ethanz :howdy, boris
(15:02:00) RMacK :patfm, as it so happens, right before we heard from you aboutblogamundo we were talking about a simpler tagging-aggregation hack
(15:02:07) RMacK :using del.icio.us and bloglines
(15:02:13) RMacK :much more primitive
(15:02:20) ethanz :but we weren’t feeling a lot of love from our contributors on theidea…
(15:02:34) patfm :i remember there was a del project along the lines of tags likegv_portuguese, or gv_brazil, or something like that
(15:02:37) ethanz :mostly because they’re already overwhelmed by the work they’re doingfor us…
(15:02:49) RMacK :yeah but the idea wasnt to ask the contributors to do it
(15:03:00) RMacK :anyway, just to explain it briefly to get that out of the way
(15:03:09) RMacK :here is the bloglines aggregator:
(15:03:09) RMacK :http://www.bloglines.com/public/GVOtranslation
(15:03:09) patfm :right. what i’m hoping to do is build something which will say “here’sa link to a translation of this post, or where you can do a translationquickly”
(15:03:25) RMacK :as it so happens, a taiwanese guy was translating gv posts into chinese
(15:03:31) BorisAnthony :but patfm I very much want to chat with you about all this sooon
(15:03:33) patfm :RMacK: ah neat, i hadn’t seen that
(15:03:39) RMacK :so i tagged them in delicious with “GVO-Chinese”
(15:03:50) RMacK :and then just fed an rss feed of that tag into a bloglines folder
(15:03:52) patfm :BorisAnthony: great, here’s our new blog: http://blogamundo.net/dev/
(15:04:06) RMacK :(yeah you havent seen it bcz i just put it together last week)
(15:04:10) patfm :ah, hehe
(15:04:14) RMacK :my thought was …
(15:04:28) RMacK :that we could let it be known that if people want to translate gv postson their own blogs
(15:04:49) patfm :right, like a request queue
(15:04:50) RMacK :to whatever extent they want, they could just tag the post withGVO-[insert language here]
(15:04:57) BorisAnthony :(argh.. my IRC client cranks my CPU usage to 90%)
(15:04:58) RMacK :and it would automatically go into that bloglines aggregator
(15:05:51) ethanz :there’s another possible step, which would involve feeding these queuedposts through a computer-automated translator for a quick and dirtytranslation which could get improved by a human translator…
(15:06:00) RMacK :true
(15:06:10) patfm :right, but that only applies in the case where machine translationexists for the language
(15:06:19) RMacK :as long as the machine-translated posts are marked very clearly asmachine translated only
(15:06:32) RMacK :so we dont create international incidents with bad machine translations
(15:06:58) RMacK :anyway, patfm, we had discussed this tagging/aggregator thing lastmonday, then i think
(15:07:01) RMacK :you sent your email tuesday
(15:07:07) patfm :there was a discussion between tim oren and kevin marks about someideas for designing a “microformat” that would mark up translationmetadata, including whether it were mt or not
(15:07:08) RMacK :perfect timing
(15:07:08) ethanz :absolutely. I see it as an open question whether or not we want toinvolve MT at all
(15:07:36) RMacK :anyway, to me it looks like we have 2 questions:
(15:08:02) RMacK :1- what is the best “hack” or combinations of hacks that can enablepeople to get started with translation using existing tools..
(15:08:21) RMacK :and 2 - what better system/set of tools can be built (enter blogamundo)
(15:08:39) ethanz :the microformat idea isn’t a bad one. would be great to have ameta-document for a blogpost that knew whether or not anyone wanted ittranslated, if translations existed and where they would live…
(15:08:49) RMacK :and how do we make sure that hack #1 enables us to transition easily tosolution #2
(15:09:09) patfm :right. so what our initial idea for blogamundo is is combination of anaggregator and “CAT” editor
(15:09:37) patfm :and the more i think about it the more i realize we’re going to need toincorporate some form of translation request queue as well
(15:09:52) patfm :i was trying to look at it from the poitn of view of a gv roundup
(15:10:00) patfm :suppose there is a roundup of, say, persian blogs
(15:10:21) patfm :is there some unobtrusive way to have some kind of link or somethingthat would indicate a translation? or that one was needed?
(15:10:38) RMacK :hmm.
(15:10:55) RMacK :are you envisioning that this persian roundup is starting out inenglish or persian?
(15:11:26) ethanz :let’s deal with arabic, since we already have ahmed, who regularlylinks to untranslated posts
(15:11:30) patfm :imagine a sentence in the (english) roundup like this: “Hoder wastalking about foo (persianlink here)”
(15:11:35) BorisAnthony :pllleeeeeaaaaase look into real web standards solutions developpedspecifically for multilanguage support before even thinking of this“microformat” sillyness!
(15:11:35) patfm :ah okay, arabic
(15:11:39) ethanz :he rounds up a bunch of arabic posts are points to the untranslatedones…
(15:11:52) BorisAnthony :the W3C has everything you need
(15:11:55) ethanz :how do we indicate that someone would like full translations of thoseposts
(15:12:00) ethanz :boris - fill us in.,..
(15:12:08) BorisAnthony :i can;t ruight now…
(15:12:09) BorisAnthony ::(((
(15:12:18) RMacK :right. i think ethan is talking about something different than whatpatfm was talking about.
(15:12:18) patfm :BorisAnthony: i’d love to talk to you
(15:12:34) BorisAnthony :patfm: i promise! very soon!
(15:12:39) RMacK :and i was talking about something different from either, when it comesto what is being translated into what by whom for what purpose
(15:12:48) BorisAnthony :i got stuck with a deadline I have to complete in the next 1-2 hours :(
(15:13:05) RMacK :so maybe before we even get into the technical details of translationwe work out what kinds of translations we are talking about
(15:13:08) patfm :no worries. i’ll be lurking in #blogamundo as well
(15:13:30) patfm :well the original thing that got me thinking aobut a possibleconnection to gv were those “dead end” links
(15:13:34) patfm :well, dead ends for me
(15:13:52) RMacK :ok gotcha.
(15:13:55) patfm :in the roundups. i would often have this feeling of really wanting toread the post on the other end of a particular link
(15:14:07) RMacK :then thats an entirely different issue than what the bloglinesaggregator was looking to solve anyway
(15:14:15) patfm :so if there were some way that i could say “i wish someone wouldtranslate this into (in my case) english…”
(15:14:22) RMacK :but which is a good problem that needs solving
(15:14:59) RMacK :i think we also have the problem of the people writing the posts notexplaining enough about what they’re linking to
(15:15:07) patfm :well, i think it’s related in a way; the main hump here seems to be amatter of letting people who can translate know there’s a place theycan do it and announce it
(15:15:12) RMacK :which is an editorial problem
(15:15:27) RMacK :right.
(15:15:41) ethanz :It would be nice if our editors/contributors could link to a post inanother language, explain a little bit what it’s about, and essentiallyrequest that someone else translate that linked post…
(15:15:51) RMacK :anyway, just to clarify in my own head, seems to me we have 2 differentkinds of translation we’re talking about:
(15:15:59) patfm :right. that’s one difference between gv & bam ; i think gv has moreof an editorial or at least organizational role. bam is intended to bean unedited channel which can be used for any purpose
(15:16:18) RMacK :for gv, anyway, there are 2 kinds of translation issues:
(15:16:22) patfm :(er, bam is what jonas & refer to blogamundo as, heh)
(15:16:38) RMacK :1. translation of actual GV posts into other languages (which is whatthe taiwanese blogger is doing)
(15:17:03) RMacK :and 2. translation of non-english blog posts we are linking to, intoenglish.
(15:17:35) patfm :ah i see. have yuo had interest in gv -> other languages besidestaiwanese?
(15:17:39) RMacK :the del.icio.us/bloglines hack i was playing with solves 1 but not 2
(15:17:48) RMacK :the taiwanese guy is writing in chinese
(15:18:02) patfm :ah whoops, my bad
(15:18:05) ethanz :very interested in both 1 and 2, pat - the question is just one ofprioritization
(15:18:08) RMacK :i have heard of people translating the odd post into french or spanish
(15:18:20) ethanz :suspect that any system you’re coming up with would actually be usefulfor both purposes
(15:18:29) RMacK :ethanz, i agree.
(15:18:30) patfm :so one thing abuot bam is that it will be open to translation betweenany language pairs
(15:18:45) RMacK :but for it to be useful for people to find stuff the output would wantto be separated out.
(15:18:55) patfm :and we’re going to have tags, and later perhaps projects (sort of likeflickr pools, i guess) where people could organize work
(15:19:21) RMacK :cool.
(15:19:31) RMacK :patfm how many people are involved w/ your project?
(15:19:41) patfm :so you cuold imagine this taiwanese blogger starting a project on bam,and then other taiwanese bloggers could contribute
(15:20:11) patfm :so far it’s jonas and myself, and my brother is advising
(15:20:16) ethanz :I think that’s the exciting possibility - that people could have acentral place for a distributed translation project…
(15:20:19) patfm :chris waigl did some great design work on the blog
(15:20:29) RMacK :right. but there would be 2 possible projects: 1 - translating GVOenglish posts into chinese, and 2 - translating the chinese stuff welink to into english. correct?
(15:20:38) patfm :right, separate projects
(15:20:57) RMacK :so the output would flow into separate places
(15:21:08) patfm :so one could imagine, for instance, a tag or a project “taishi”, andpeople could just translate cn -> en there for instance
(15:21:17) patfm :and another tag/project “globalvoices”
(15:21:32) RMacK :right.
(15:21:34) patfm :right. we’re still in flux aobut the url structure, as i babbled a bitin that email ot the list
(15:21:41) ethanz :right. and there are subprojects to globalvoices - en->esp, en-fran,etc.
(15:21:50) RMacK :in the case of the first project you mention, it may or may not relateto stuf that gv actually linked to. correct?
(15:22:05) RMacK :(question directed at patfm)
(15:22:10) patfm :but something like blogamundo.com/tags/tashi/cn/en ;blogamundo.com/tags/globalvoices/en/cn
(15:22:43) RMacK :patfm that makes sense
(15:22:46) patfm :well blogamundo will be an open place for translation, yes. it won’t beexclusively gv content. i just think that there could be a nicesymbiosis ofr a while
(15:22:55) RMacK :absolutely i agree
(15:23:05) RMacK :big potential for synergy
(15:23:23) ethanz :something I think we’d be very psyched not to build ourselves
(15:23:27) patfm :it’s fun to think of applications, there really are a lot. ndesanjomailed me and seemed interested in my hypothetical idea in the list
(15:23:47) RMacK :as long as boris is happy with the tech and feels that it integrateswell with the system he’s building for us, blogamundo could be where gvtranslation work “lives”
(15:23:48) patfm :yeah, it’s keeping us busy full time
(15:23:50) ethanz :possible that we’ll have some funding for translators in the future -if the toolbase goes somewhere, I can certainly imagine usexperimenting with them…
(15:24:29) RMacK :so i have a question.
(15:24:37) patfm :one of the reasons i started that blog is to try to get into contactwith professional translators and discuss what they think about thewhole system, what they would expect from a commercial option, etc
(15:24:48) ethanz :hey ndesanjo…
(15:24:56) patfm :(i have a lot to learn there)
(15:24:57) RMacK :hey ndsanjo, good to see you last week!
(15:24:58) patfm :hi ndesanjo!
(15:25:05) ndesanjo :hey everybody
(15:25:08) RMacK :so anyway, patfm I have a question
(15:25:21) ndesanjo :it was great rebecca
(15:25:28) patfm :shoot
(15:26:11) RMacK :so anyway, until pafm’s thing is up and running, and if we want tostart experimenting with tagging/aggregation of whatever volunteertranslation people want to do on their own blogs,
(15:26:41) RMacK :should we agree upon a set of del.icio.us tags, etc.,
(15:26:54) RMacK :that would be most easily portable, plug-in-able, etc.?
(15:27:37) patfm :it wouldn’t hurt, as a way to bootstrap translation requests
(15:27:40) RMacK :i think it would be interesting to start publicizing the fact that ifpeople translate GVO posts, we will aggregate everything tagged acertain way
(15:28:09) patfm :hmm yeah that’s an interesting idea
(15:28:10) RMacK :and that if people want to request translation for certain non-englishposts, they can tag it a certain way,
(15:28:19) RMacK :and it will go into another aggregator that takes in the request-tags.
(15:28:29) patfm :do you have any idea of how many gv readers use delicious, justanecdotally?
(15:28:43) RMacK :probably a real minority
(15:28:51) ethanz :I think it’s totally worth trying, RMacK… I just don’t know how manyof our potential translators are delicious/tag savvy…
(15:29:04) RMacK :ethanz i agree
(15:29:08) RMacK :that’s the problem.
(15:29:23) patfm :one thing that jonas and i had discussed was the possibility ofgenerating bookmarklets on a per-language basis
(15:29:29) RMacK :if somebody would like to propose a hack using existing tools thatdoesnt require tagging, pls feel free
(15:29:57) patfm :that would add translation requests to a request queue, either ondelicious or perhaps on blogamundo. but we hadn’t worked out details
(15:30:09) RMacK :patfm, yeah.
(15:30:26) patfm :that would have the advantage of just being “drag this request buttonto your toolbar”
(15:30:34) RMacK :i think what we’re trying to figure out first is what we can startdoing before you get blogamundo up and running
(15:30:49) RMacK :which would integrate most easily once blogamundo gets going
(15:31:12) BorisAnthony :yup
(15:31:18) patfm :gotcha
(15:32:50) patfm :so RMacK, would they be language tags, or something more general like“translationrequest”
(15:33:01) patfm :or a set of tags, maybe like:
(15:33:30) patfm :{from:english to:português translateme}
(15:34:20) RMacK :yeah probably something like the latter
(15:34:47) patfm :one thing i’ve wonder about is how familiar people are with the isolanguage abbreviations
(15:35:11) ethanz :even if they’re unfamiliar, it’s a standard, and standards are a usefulthing in cases like this
(15:35:27) patfm :for instance, would most kiswahili speakers recognize a link or tagthat looked like [sw] or [swa]
(15:35:46) ethanz :would be great to agree on a simple tag standard that used Isoappreviations and supported requested translations and completedtranslations
(15:35:47) RMacK :good question.
(15:36:03) RMacK :that’s a good idea ethanz
(15:36:10) patfm :i ask because one little convention that occurred to me would be tohave links that looked like: “link to article” [sw] [»en] or some such
(15:36:34) RMacK :and useful for patfm to maybe put together a “standard tags page” assomething like that will be needed for blogamundo anyway
(15:36:55) RMacK :then we can use the same standards in tagging stuff before blogamundogets started,
(15:37:10) RMacK :which means that it can all plug right in when b-mundo gets up andrunning
(15:37:22) ethanz :hi sokari…
(15:37:26) patfm :that’s a good idea. we could make a little database that served a “findyour language tags” sort of purpose
(15:37:42) patfm :hey sokari, i’ve seen you on the intarweb!
(15:38:05) RMacK :yeah, patfm, good idea.
(15:38:11) RMacK :or even a wiki page for starters
(15:38:14) ethanz :I think the path might be to let Pat and crew think through the taggingissue for a few days. when they propose a standard, we can pitch thatto our collaborators. that would put us on a common standard so thatwhen blogamundo is live, we could port our efforts over to it easily.is that a good set of next steps?
(15:38:19) RMacK :which wld get turned into a database
(15:38:37) RMacK :ethanz that sounds like a most excellent plan
(15:38:43) patfm :that sounds good to me
(15:39:02) RMacK :i had not yet publicized our del.icio.us/bloglines hack and will not doso till our tagging standards get sorted
(15:39:04) patfm :i’d also like to talk to BorisAnthony about his ideas in the near term
(15:39:31) RMacK :patfm good point
(15:39:55) ethanz :I think Boris will be able to fill RMack and me in a bit tomorrow. andyou guys should find some time to talk over the next week, Pat…
(15:39:55) RMacK :you around tomorrow patfm?
(15:40:14) RMacK :yeah, i think ethan’s right
(15:40:16) patfm :i’m always around :)
(15:40:28) RMacK :we’ll hunt you down on irc
(15:40:39) RMacK :boris will be here tomorrow but might get caught up in meetings all day
(15:40:40) patfm :we’ve opened #blogamundo as well, so i’ll be in there pretty much allthe time now
(15:40:41) RMacK ::-)
(15:40:45) RMacK :awesome
(15:40:46) patfm :great!
(15:40:49) RMacK :so we know where to find you
(15:41:12) RMacK :meanwhile do you want us to blog about blogamundo?
(15:41:22) RMacK :to help you find people who want to help?
(15:41:29) patfm :hmm good question
(15:41:38) RMacK :or do you want to keep it quieter for a while
(15:41:49) patfm :let’s hold off until the tags are settled
(15:41:56) RMacK :yeah that makes sense
(15:41:56) patfm :so a week or so
(15:42:12) RMacK :cool
(15:42:14) patfm :we want to be able to give people something to do right away
(15:42:34) RMacK :yeah i think giving people something to do right away using existingtools can be useful
(15:42:51) RMacK :bcz then you’ll see patterns of behavior and issues emerge that youcant anticipate
(15:43:00) RMacK :and patterns of use, patterns of demand, etc
(15:43:02) ethanz :I think that’s right - spend a bunch of time thinking through the tags,then promote that as a useful first step…
(15:43:13) patfm :right. we expect user requests to be a major force in shpaing what bamdoes.
(15:43:20) RMacK :you never know who is going to seize upon the whole thing and getreally into it, and for what reason
(15:43:38) RMacK :user-shaped tools ++++ :-)
(15:44:03) patfm :right, exactly. the thing we’re kvetching over most is url structure,whether to use groups *and* tags, that sort of thing
(15:44:12) patfm :because we want it to feel maximally open to user goals
(15:44:48) patfm :we just aren’t sure whta those goals will be yet. actually that’s why ithink gv is a great first relationship; you guys already ahve a clearmission
(15:45:12) patfm :and the roundups function already, bam can just help add to those overtime
(15:45:18) RMacK :yeah its a great relationship :-)
(15:45:31) RMacK :tremendous synergy potential!
(15:45:44) RMacK :what to do about translation has been a huge issue lately
(15:45:51) RMacK :patfm remind me where are you based now?
(15:46:01) patfm :i’m in silver spring maryland, right on the border with dc
(15:46:18) RMacK :thats right
(15:46:23) patfm :my background is actually in linguistics
(15:46:29) RMacK :gotcha
(15:46:49) RMacK :now its all coming back!
(15:47:01) RMacK :alewife is such a new england kind of name
(15:47:11) chryss :(sorry, my wifi has become very crappy… hi all)
(15:47:51) patfm :i love it up there. visited boston once… dunkin donuts coffee onevery corner, my kind of place :)
(15:47:52) ethanz :hi chryss
(15:47:52) RMacK :really really excited to work out this translation thing
(15:48:11) RMacK :back to your tags vs. groups thing…
(15:48:46) patfm :ah, jonas just put up some screenshots http://blogamundo.net/ss/
(15:48:51) RMacK :the main thing should probably be that every piece of public contenthas its own permalink
(15:48:57) patfm :there’s no layout yet but you can get an idea of the functionality
(15:49:34) RMacK :cool
(15:49:53) RMacK :brb
(15:51:04) patfm :ideally the requests would include a specific permalink as well, tho iworry that some folks will just link a blog url
(15:51:14) patfm :not much to be done about that problem
(15:51:51) patfm :by the way, the editor in that screenshot is VERY preliminary, heheh
(15:52:30) patfm :it will actually handle translation on a sentence-by-sentence level.here’s one mockup we put together http://ruphus.com/stash/editor.html
(15:52:52) jgalvez :extremely. expect obscene amounts of *cough* web2.0 *cough* ajax*cough* in it
(15:53:06) patfm :heheh, hey jonas
(15:53:21) jgalvez :** the plumber emerges
(15:53:40) patfm :hehe
(15:53:59) RMacK :like the mockup
(15:54:06) RMacK :this is a very cool project
(15:54:42) patfm :we’ve drifted away from the dishsoap yellows, however :)
(15:56:12) patfm :would you guys mind if we posted this chat somewhere? i know chadfowler had wanted to take part but he is recovering from a redeye
(15:56:59) RMacK :sure.
(15:57:02) BorisAnthony ::D
(15:57:02) RMacK :by all means
(15:57:51) patfm :well thanks tons for setting this up rebecca! great talking to you guys
(15:58:06) RMacK :no worries at all. really glad we’ve started this conversation.
(15:58:09) RMacK :lets keep talking.
(15:58:20) RMacK :let us know when you’ve thought more about what you want to do with thetagging system
(15:58:36) RMacK :and we’ll remind boris to talk to you when he gets a chance!
(15:58:37) ethanz :yeah, I think our next checkpoint is when you guys have a proposedtagging standard
(15:58:38) patfm :let’s see, contact stuff: my email is pathall@gmail.com or pat@blogamundo.com, the dev blogis http://blogamundo.net/dev
(15:58:44) ethanz :we can do some pr stuff around it then
(15:59:04) patfm :awesome
(15:59:06) RMacK :ethanz is right
(15:59:16) RMacK :fab-a-mundo
(15:59:32) patfm :and by the way, if you guys ever need some more odd css hackery aroundgv, look up chris waigl, she cranked that whole blog out in 2 days
(15:59:46) patfm :hehe. okay, i’m audi 5000, thanks again!