Boston SMW users meeting

From Tetherless World Wiki

Jump to: navigation, search

The first meeting of Semantic MediaWiki users is scheduled to be held in Boston, Massachusetts, USA on Saturday, November 22, 2008, with a "tech day" on Sunday, November 23.

Contents

Overview

The main meeting, on Saturday, will start at 10:30 AM and will cover many aspects of SMW and its related extensions, including technical, usability, training, "best practices", marketing and business-development issues. The Sunday, "tech day" meeting will also start at 10:30 AM, and will cover in-depth technical aspects of some proposed additions to the SMW system.

Minutes

Notes by Li

http://tw.rpi.edu/wiki/Boston_SMW_user_meeting_2008-11-22

Notes By Shangguan

Global outreach:

  • what is the targeted audiences of SMW?
    • IT industry
    • Academic
    • user/developer
    • others
  • how to get Wikipedia to adopt SMW
    • some progress already
    • scalability issues -- RPI' initial experiment using DBPedia
  • semantic wiki forum by ontoprise


"LexWiki Framework & Use Cases"by Guoqian Jiang @Mayo Clinics

(I miss this talk, so just give the slide and link here)


"Semantic MediaWiki and Exhibit" by Fabian Howahl @MIT CSAIL (visiting undergrad.)

  • integration of SMW and Exhibit
    • Exhibit: more like a faceted knowledge visualization tool (web-based)
  • SMW's isolated from each other
    • solution: cross wiki queries, parsing .jason file which contains query results (looks like Evan's work)


"Future Developments"by Yaron Koren

Notes by Evan

  • SMW Developer/User Meeting
    • Presented Ask Many Extension to allow for inter-wiki querying
    • Extension seemed well received, many people were looking for such a tool
    • A lot of curiosity regarding ACL and other security techniques for protecting triples
  • Competition
    • Fabian Howahl
    • Visiting MIT from Karlsruhe
    • Wants to use Exhibit to display interwiki content

External Notes

Yaron Koren

http://sourceforge.net/mailarchive/forum.php?thread_name=55dbda5c0811241234m36657260h2397dd422eea42ba%40mail.gmail.com&forum_name=semediawiki-user

Here's my report on the first-ever Semantic MediaWiki users meeting, which was held in Cambridge, Massachusetts over the weekend, at the offices of the Monitor Group. The meeting was divided into two days: the Saturday session was an all-day event that featured discussions and presentations on a wide variety of SMW-related topics, and the Sunday session was a half-day event that focused on two topics: in-depth discussion of possible additions to the SMW and its spinoff extensions, and ideas about building and strengthening the SMW community, including thoughts about future meetings. There were about 30 attendees for the Saturday session and 10 for the Sunday session, and both numbers were significantly than I expected when we first started talking about the meeting. There were attendees from both Germany and various parts of the United States, and among the attendees were representatives of consulting companies who implement SMW-based solutions (Monitor, Ontoprise, mindXchange, SAIC), university groups where SMW-based research is done (RPI, MIT, "Karlsruhe Institute of Technology", i.e. the University of Karlsruhe), and large-scale websites built using SMW (caBIG Knowledge Center, FamilySearch, SNPedia, Referata), among others. Also included were "special surprise guests" Markus Krotzsch and Denny Vrandecic, who phoned in via Skype for the Saturday morning sessions.

Overall, I think the meeting was a success - many connections were made between people working on similar problems, or encountering the same obstacles, and there was a great deal of information exchanged. And the facilities and equipment were superb, thanks to some great work from the Monitor team.

There were too many thoughts raised during the meeting to try to summarize all of them, but here's my attempt to list a few of the recurring ideas that came up:

- There should be more communication. This is true for nearly everyone - groups and individual developers who create features or extensions should announce what they're doing before they start developing, not after it's done, and should solicit feedback more; consulting companies should communicate more about the projects they're working on; etc. This mailing list and the SMW Forum (http://smwforum.ontoprise.com

) were raised as the two best places to do that, although of course any communication is better than none.

- There should be more documentation. That includes codifying "best practices" and use cases, and expanding help pages. On that note, some pleas were made to re-open registration on semantic-mediawiki.org (sorry, Markus and Denny, I'm just the messenger). And more generally, a need for tutorials was expressed, including online screencast videos, PDF documents and maybe even published books.

- Many new features were requested. The number one request seemed to be for access control through SMW - letting administrators close off parts of a wiki for viewing (and, less importantly, for editing), using semantic properties - and making sure that SMW inline queries respect that and hid results that users shouldn't see. Thankfully, according to Ontoprise they're working on implementing such a feature, although it's still in the planning stage. Also popular was workflow, which is tied in with access control and which involves, most notably, various user groups being notified when specific changes are made to the data; and what I call "true n-ary relations", which is the ability to define free-form internal subjects within a page, each with their own set of properties.

- Finally, there was discussion about the need for future user meetings. It was suggested that semantic-related conferences, such as SemTech and ISWC among many others, could include either official or unofficial SMW-specific meetings. In addition, it was tentatively decided that there should be two dedicated SMW user meetings such as this one every year, most likely in the fall and spring, and most likely alternating between the United States and Europe.

So that's it for my notes. You can read more about the discussions, view slide presentations and see some photos, from the two links at the top of this page:

http://smw.referata.com/wiki/SMW_users_meeting

The minutes are still being added to, but hopefully what's there provides a good sense of how the meeting went.

Joel Natividad

SMW UGM-1 Nov 24 Livescribe Session

http://www.livescribe.com/cgi-bin/WebObjects/LDApp.woa/wa/MLSOverviewPage?sid=4Q5k29gnHmMm

Chat History

Chat History with SMW User Meeting (#kroetzsch/$yaron.koren;8ef16a8607657b69)

Created on 2008-11-24 12:55:46.

2008-11-22
Denny Vrandečić
11:19:14
hi
Harold Solbrig
11:19:24
THNX
Guoqian Jiang
11:19:31
thanks
liding2006
11:19:58
How to get SMW in use
  • SMW on stick (for personal use)
  • online SMW show cases (for group/corperate use)
  • Inject to wikipedia
Denny Vrandečić
11:20:08
everyone can pull in new users intothe chat
Harold Solbrig
11:20:19
What do I get from SMW that I don't get from plain MW?
Harold Solbrig
11:20:27
This is part of the selling...
Marc Feickert
11:20:34
a capital S ;)
Harold Solbrig
11:20:44
lol
Denny Vrandečić
11:20:56
cool maps with data showing on it
Markus Krötzsch
11:20:58
I guess the S frightens people.
Denny Vrandečić
11:21:12
and less work for getting more consistent data
Denny Vrandečić
11:21:24
a wiki that can also be read by a machine
Harold Solbrig
11:21:45
Maybe - from our perspective, the ability to export "knowledge" in RDF is a selling point, but the format and semantics need discussion.
Marc Feickert
11:22:06
That'sa big plus that we are pursuing... integration of a genealogy research wiki with the main FanilySearch website
Markus Krötzsch
11:22:08
@Denny: not sure if this is convincing to normal users; I find them having a hard time understanding this "machine-readability" things
Harold Solbrig
11:22:09
Playing devil's advocate - doesn't WIF do that as well? The Wiki API?
Denny Vrandečić
11:22:32
i wouldnt use rdf as a selling point for end users, i'd call it "machine readable data exports in standard formats"
Harold Solbrig
11:22:50
I'd like to propose one - the ability to import, represent and export formal models (e.g. UML, XML Schema, etc.)
Marc Feickert
11:22:52
MRDEISF
Markus Krötzsch
11:22:57
In fact, we are already supporting many other formats besides OWL/RDF
Markus Krötzsch
11:23:54
waves at Fabian
liding2006
11:24:12
Yaron: How people found out Semantic Wiki
  • found from google,
Harold Solbrig
11:24:18
MRDEISF?
Denny Vrandečić
11:24:29
what does mrdeisf mean?
Marc Feickert
11:24:41
machine readable data exports in standard formats
Marc Feickert
11:24:57
Sorry, I'll keep the bad chatter down.
Guoqian Jiang
11:25:01
Harold, could you speak out what brought your idea for SMW use?
Markus Krötzsch
11:25:05
oh, right, good there is finally one acronmy to rule them all
Harold Solbrig
11:25:10
Will do
Denny Vrandečić
11:25:11
 :)
Denny Vrandečić
11:26:19
who is speaking? sorry missed that
Marc Feickert
11:27:03
I think oneof the main selling points for many organizations would be if SMW could be presented with a default intuitive UI. Yes, we can design our own, which is fine, but if you want it to spread to a different audience, it needs to come to their level
Markus Krötzsch
11:27:30
alas, this could be said for MediaWiki already ...
Marc Feickert
11:27:38
Indeed
Marc Feickert
11:27:52
That's where i hope the SMW+ package can really help
Markus Krötzsch
11:28:12
yes, there are good UI improvements there, also for parts of MW
liding2006
11:28:29
* browse Mediawiki extension
  • Jie: build semantic wiki from sratch and but not succesful adopted, then came across SMW
  • Li: not convinced by performance initally in 2006, in 2008, get referral by Jie Bao through Social network
  • reading presentation by famous presentation: TBL's future Web 3.0 talk
Markus Krötzsch
11:29:03
sure
Markus Krötzsch
11:29:12
we just mute our mic when we do not talk
Marc Feickert
11:29:15
It was what made it possible for FamilySearch to adopt SMW. We don't have a team of developers, and our audience is very unfamiliar with computers at all, let alone this
Markus Krötzsch
11:30:26
Great to see that you could get something working on that basis!
Denny Vrandečić
11:30:34
what is the url of familysearch?
Harold Solbrig
11:30:38
To work, there needs to be community support - either in terms of resources, publicity, etc or in terms of a bit of funding.
Mollie Forbes
11:30:41
wiki.familysearch.org
Marc Feickert
11:30:44
for the wiki, or the main site?
Mollie Forbes
11:31:00
our public wiki is not currently running SMW
Harold Solbrig
11:31:08
I don't mind working to get documentation put together as long as I know it will be used and bought into...
Marc Feickert
11:31:47
www.familysearch.org is the main

wiki.familysearch.org is the wiki

As Mollie said, SMW is under development currently, but you can see what the wiki is about and what the content is like

Denny Vrandečić
11:37:18
+1
Denny Vrandečić
11:37:25
(who is talking?)
baojie_iowa
11:37:41
Harold
Denny Vrandečić
11:38:32
fabian, jie, li, did you already present your work in federating smw's?
Marc Feickert
11:38:55
They have not
Denny Vrandečić
11:39:34
handraise
Denny Vrandečić
11:39:40
on sparql endpoint
Marc Feickert
11:39:41
(y)
Marc Feickert
11:39:52
(rofl)
Harold Solbrig
11:40:15
I guess that they didn't understand what that meant?
Marc Feickert
11:40:22
They figured it out
Denny Vrandečić
11:40:25
no :)
Denny Vrandečić
11:40:28
i'll type it
Markus Krötzsch
11:40:30
(can you actually hear us?)
Harold Solbrig
11:40:36
Not very well.
Harold Solbrig
11:40:48
They're going to try to connect to the speakers on the break.
Denny Vrandečić
11:40:52
the sparql endpoint currently is based on RAP, which is very badly supported
Markus Krötzsch
11:41:07
we will be aorund only until lunch break anyway
Denny Vrandečić
11:41:08
during iswc i have talked to benjamin nowack, who has won the semantic web challenge
Yaron Koren
11:41:10
Yeah, we're working on it.
Yaron Koren
11:41:19
(The speakers and microphone)
Michael Cariaso
11:41:21
may be room for improvement on RAP, but I'll chime in it works great for me, and has been a real game changer
Denny Vrandečić
11:41:32
he has a php rdf store written by himself
Markus Krötzsch
11:41:36
We should be more explicit: RAP is abandonded.
Denny Vrandečić
11:41:39
arc2 which seems to work much better
Denny Vrandečić
11:41:59
benjamin will work hard on integrate arc2 and smw much better than it was possible for us with rap
Denny Vrandečić
11:42:07
so we want to replace rap with arc2 in short term
Denny Vrandečić
11:42:20
for the outside world, nothing should actually change
Denny Vrandečić
11:42:34
michael, so for you it should work just as before, just better :)
Denny Vrandečić
11:43:00
michael, would you mind being a tester on the new implementation?
Michael Cariaso
11:45:33
well, I need my machine to stay up
Michael Cariaso
11:45:38
there is steady traffic
Michael Cariaso
11:46:02
but if its fairly easy to swap back and forth
Markus Krötzsch
11:46:22
yes
Michael Cariaso
11:46:29
then yes I'd be very happy to
Markus Krötzsch
11:46:41
ew hear you well
Denny Vrandečić
11:46:42
we hear you
Markus Krötzsch
11:46:56
can you hear us?
Michael Cariaso
11:47:02
I should make this fairly clear, the site I host is www.SNPedia.com
Denny Vrandečić
11:47:18
oh wow, snpedia is really a coooool idea!
Michael Cariaso
11:47:53
there is a sparql endpoint at /rdfapi-php/netapi/snpediamodel
Michael Cariaso
11:48:09
my thanks
Denny Vrandečić
11:48:31
ok, i'll jot you down and connect you with benjamin as soon as we start with the new sparql implementation
Denny Vrandečić
11:48:40
it should be quite an improvement on the rap based one
Michael Cariaso
11:49:33
I'd be very happy to play tester, your code has served me very well.
Denny Vrandečić
11:49:41
cool, thx :)
Denny Vrandečić
11:50:06
i hope to get a dna screening myself in mid term, so i can connect it to snpedia :)
Michael Cariaso
11:50:25
k, when you do
Michael Cariaso
11:50:33
I can definitely help a bit
Marc Feickert
11:51:01
Hey, I'd really like to talk to those of you who work with the ontologies. I need some pointers and advice on our own
Michael Cariaso
11:51:04
the whole point of SNPedia, is so I can build reports like http://www.snpedia.com/files/promethease/outputs/?C=S;O=D
Marc Feickert
11:51:10
be it in a breakout or dinner, whatever
Denny Vrandečić
11:51:12
thx, i'll come back to you, michael ;)
Michael Cariaso
11:51:17
ttfn
Denny Vrandečić
11:52:01
marc, in what sense?
Markus Krötzsch
11:56:06
Are there links to slides for presentations?
Marc Feickert
11:56:58
http://tw.rpi.edu/wiki/Images:Semantic_Wiki%40RPI.ppt
Markus Krötzsch
11:57:03
great, thanks
Denny Vrandečić
11:57:09
thx
Markus Krötzsch
11:57:36
the real link is [images/8/8a/Semantic_Wiki%40RPI.ppt http://tw.rpi.edu/wiki.tw/images/8/8a/Semantic_Wiki%40RPI.ppt]
Guoqian Jiang
11:57:38
can I send my slides about lexwiki to this group
Yaron Koren
11:57:56
Yeah, go ahead.
Marc Feickert
11:58:05
Ontology advice: I am a budding ontology engineer, and I don't know or am having trouble understanding some of the tricks and methods that are particular to SMW. So I want to learn more from experience people what some good practices and tricks are, and also get some opinions on how to adjust and improve what I have already built
Guoqian Jiang
11:58:05
sent file "LexWiki_SMWUsersMeeting_20081122.ppt" to members of this chat<files alt=""><file size="3296256" index="0">LexWiki_SMWUsersMeeting_20081122.ppt</file></files>
Marc Feickert
11:58:22
That's in answer to the question above
Yaron Koren
11:58:45
Oh, the file itself... I don't know if that worked.
Guoqian Jiang
11:59:29
well, please let me if it works or not
Denny Vrandečić
11:59:56
marc: ah, i understand. if you have a breakout, i hope you'll take notes :)
Denny Vrandečić
12:00:03
we have lost audio
Markus Krötzsch
12:00:18
Yaron, could you call us in again?
Michael Cariaso
12:00:32
re: How people found out Semantic Wiki . I was building lots of custom apps based on top of mediawiki, discovered Halo first, then SMW from http://www.mediawiki.org/wiki/Extension_Matrix
Yaron Koren
12:00:55
Calling you now.
Markus Krötzsch
12:01:02
call Denny please
Yaron Koren
12:01:14
Oh.
Marc Feickert
12:01:16
Denny: You'd better believe I will!!
Denny Vrandečić
12:01:26
thx
Markus Krötzsch
12:01:30
ok, back
Denny Vrandečić
12:01:30
hearing again
Markus Krötzsch
12:01:45
Which slide are we on?
robertrfreimuth
12:01:46
Marc: The caBIG Vocabulary Knowledge Center is attempting to create docs and training material for use cases like that. I don't think we have enough ready right now (and it's certainly not organized very well), but more will be coming soon. See https://cabig-kc.nci.nih.gov/Vocab/KC.
Harold Solbrig
12:01:58
I found SMW through Eric Mueller (sp?) - he'd been working on Wiktionary, but we'd all agreed that it didn't have the degree of formality needed. I searched for his name occassionaly and found a co-reference.
robertrfreimuth
12:02:25
I'll be giving a brief overview of the VKC and how SMW is being used for cancer research in this session, during Guoqian's block on LexWiki.
Marc Feickert
12:02:29
Thanks, Robert. I'll give it a look.
Guoqian Jiang
12:04:34
it seems hard to get direct file transfer through skype
Marc Feickert
12:04:51
you get relayed like mad if you're not on the same network
Markus Krötzsch
12:04:54
yes, it is slow if you have both ends sitting inse
Yaron Koren
12:05:01
Guoqian, can you upload it to the web?
Markus Krötzsch
12:05:03
.ide company nets
Markus Krötzsch
12:06:59
All SMW extensions are usually found via http://www.mediawiki.org/wiki/Category:Semantic_MediaWiki_extensions
Yaron Koren
12:07:16
Yeah - this is all RPI's extensions.
Yaron Koren
12:07:38
Although one, "Tetherless Map", was just added to mediawiki.org last week.
Markus Krötzsch
12:07:47
yes
Markus Krötzsch
12:08:25
(obviously mediawiki.org should use SMW to improve their extension management ;-), some of those are rather outdated)
Yaron Koren
12:09:00
Hey, good idea.
Steven Forth
12:09:22
And I agree with Mark that we should default o putting things of SourceForge
Denny Vrandečić
12:09:33
the audio is getting weak
Denny Vrandečić
12:09:34
very weak
Denny Vrandečić
12:09:38
better
Yaron Koren
12:09:50
I think it all depends on which direction Li is facing. :)
Markus Krötzsch
12:09:55
yes
Denny Vrandečić
12:09:56
ah
Harold Solbrig
12:09:57
Page - SimpleExport (Extension)
Harold Solbrig
12:10:19
(If you're following the slides - he's talking about N3 format)
Markus Krötzsch
12:10:32
yes, thanks!
robertrfreimuth
12:10:34
Is the room mic directional? It looks like it is pointed into the middle of the room and away from the speaker.
Harold Solbrig
12:10:51
I think it is - can someone get Yaron's attention?
Harold Solbrig
12:11:13
Is that better?
Denny Vrandečić
12:11:16
yes
Yaron Koren
12:11:25
Oh... I guess it was using the mic.
Yaron Koren
12:11:29
Oops.
Yaron Koren
12:11:35
Thanks, guys.
Harold Solbrig
12:12:09
Mayo has been working on standardized template sets for particular purposes and XML/Template transformations.
Denny Vrandečić
12:12:26
xml/template trafos?
Joel Natividad
12:12:27
just a suggestion - what if everybody in the meeting mark up their profiles using the Semantic Form Person template on smw.referata.com. Perhaps, we can modify the template to add resources
Markus Krötzsch
12:12:46
URL?
Harold Solbrig
12:12:50
While we argue that Template / XML is (close to) isomorphic...
Joel Natividad
12:13:11
http://smw.referata.com/wiki/Form:Person
Steven Forth
12:13:25
I will mention this when there is an opening.
Denny Vrandečić
12:13:48
i'm marked up :)
Guoqian Jiang
12:14:02
I uploaded lexwiki slides into a mayo ftp server ftp://ftp.mayo.edu/pub/lexwiki/, there are two slides file there. user name: mayoftp password:your email address. please let me know if this works
Marc Feickert
12:14:54
yeah, it's good
Marc Feickert
12:15:19
download speed sux
Marc Feickert
12:15:37
26K
Guoqian Jiang
12:15:46
good
Marc Feickert
12:15:47
make that 11K
Markus Krötzsch
12:16:21
it was much faster for me
Marc Feickert
12:16:26
I need a dial-up modem to speed up the transfer!
robertrfreimuth
12:16:34
11k and decreasing? At that rate we'll just convert to binary and paste the 1's and 0's into skype.
Marc Feickert
12:16:40
Markus, you are not at Monitor
Markus Krötzsch
12:16:52
I known :-)
Marc Feickert
12:16:53
this wireless connection is a bottleneck, i think
Marc Feickert
12:16:58
(fubar)
Markus Krötzsch
12:17:07
I hope our Skype-Audio is not afected by this ...
Yaron Koren
12:17:26
Well, no need to download the slides if you're here.
Marc Feickert
12:17:30
no, that's port 80, so it appears to be better
Denny Vrandečić
12:17:31
could rpi send us an email, if they did not yet, on the details of the scalability issues?
Marc Feickert
12:17:34
port 21 is slow
Yaron Koren
12:17:56
Denny, I'll tell them.
Denny Vrandečić
12:18:01
thx
Michael Cariaso
12:18:12
111000111000111111000 LOL
Denny Vrandečić
12:18:29
audio is lossing
robertrfreimuth
12:18:33
Whups - you've got a typo there. That 1 should be a 0... ;-)
Denny Vrandečić
12:18:45
better
Steven Forth
12:18:48
Mike direction again, is it better now
Yaron Koren
12:18:52
I had to point the mic.
Markus Krötzsch
12:18:56
yes, thx
Marc Feickert
12:19:05
There are 10 people in the world... those who read binary and those who don't
Joel Natividad
12:19:22
lol
Yaron Koren
12:19:22
 :)
Denny Vrandečić
12:19:55
the problem with multilinguality in a wiki is that the wiki text is still in one language
Denny Vrandečić
12:20:02
so you can have the data being represented in multiple labels
Denny Vrandečić
12:20:13
but you cannot translate the text itself
Marc Feickert
12:20:42
Yes. You could search across multiple wikis, but not have connected translationsof pages
Denny Vrandečić
12:21:04
right. but the data could be woven together. not the textual content
Marc Feickert
12:21:24
download finished (whew)
Neil Davies
12:21:43
rotate mike
Denny Vrandečić
12:21:56
which talk is that now?
baojie_iowa
12:22:01
Evan Patton, RPI
baojie_iowa
12:22:02
http://tw.rpi.edu/wiki/Image:AskMany.pdf
Denny Vrandečić
12:22:31
thx
Harold Solbrig
12:23:38
Do you guys on the far end have this set of slides?
Harold Solbrig
12:23:47
If so, I could try to give queues...
Markus Krötzsch
12:23:53
yes, I just downloaded it
Harold Solbrig
12:24:14
He is on the AskMany Extension page - bullet 4
Markus Krötzsch
12:24:26
great, that helps!
Harold Solbrig
12:24:38
askmany xml
Harold Solbrig
12:25:23
Embedding SMW query inside raises a flag...
Denny Vrandečić
12:25:28
so does the category has to be "Person" on both wikis? what if it is "Person" on the one, and "People" on the other?
Harold Solbrig
12:25:39
... why not xquery (or more important, why not XML???)
Harold Solbrig
12:25:52
Shared model is key to this, isn't it?
Markus Krötzsch
12:26:32
Why not Special:Ask?
Neil Davies
12:26:40
if anyone wants to do screenshare
Neil Davies
12:26:42
https://webmeeting.monitor.com/smwusermeeting/
Markus Krötzsch
12:26:50
(Special:ExportRDF oviously is not for queries)
Harold Solbrig
12:26:50
Efficiency slide
Markus Krötzsch
12:26:53
thx
liding2006
12:27:08
for denny's question:
Harold Solbrig
12:27:16
Current Status slide
liding2006
12:27:24
concept mapping is not yet implmented, but being planned
Markus Krötzsch
12:27:49
that would be useful, what method do you plan to use?
Harold Solbrig
12:28:01
It is a bit of a challenge. Recommend you take a look at the "LexWiki" model as an approach...
liding2006
12:28:27
denny, evan get back to your question
liding2006
12:29:21
unlike special ask
Harold Solbrig
12:29:23
Is this a job for SMW or a distributed RDF?
Harold Solbrig
12:29:33
Future Work slide.
liding2006
12:29:36
markus, the askmany can ask manywiki site
Harold Solbrig
12:29:54
QUestions Slide
Markus Krötzsch
12:29:59
yes, but it could use Special:Ask to get the data from remote sites
Markus Krötzsch
12:30:03
or does it?
liding2006
12:30:13
yes
liding2006
12:30:31
we did use ask code to execute query on remote wiki
liding2006
12:30:52
askmany just distributed request and aggregate the returned results
liding2006
12:31:14
p.s. the updated presentation for my slide is
liding2006
12:31:15
http://tw.rpi.edu/wiki/Image:Semantic_Wiki%40RPI.pdf
Markus Krötzsch
12:31:16
Special:Ask could easily be extended to support RDF export
Denny Vrandečić
12:31:17
cool
Yaron Koren
12:31:31
Or maybe XML? Or bare HTML?
Markus Krötzsch
12:31:50
or JSON ... wait for Fabian's talk
liding2006
12:31:51
evan did not export RDF, but just the results encoded in HTML generated by Ask
Denny Vrandečić
12:32:15
oh, how did he mash the results then?
liding2006
12:32:33
I have to ask evan
liding2006
12:32:51
we've made many simplication to get it work first
Yaron Koren
12:32:51
I think AskMany mashes the results directly.
Denny Vrandečić
12:33:15
markG is right, it would be really good to solve that security problem, but it is really hard to solve it
liding2006
12:33:25
yes
Marc Feickert
12:33:29
It's a big one for us.
liding2006
12:33:32
right now, ask can query any triples,
Harold Solbrig
12:33:37
Are there plans for an official triple-store as a back-end option (vs. import/export?)
Markus Krötzsch
12:33:39
data formatting could also be very tricky in this approach, because the HTML is user-readable but not necessarily machine-readable (e.g. with other language settings, you get other HTML for the same data)
liding2006
12:33:44
that stop some our students publishing semantic content
Marc Feickert
12:33:47
We don't want people like my grandmother playing with the core ontology
Daniel Hansch
12:33:54
li: who is the contact person at RPI for ACL on triples?
Harold Solbrig
12:34:05
We would like to be able to query (and update!) the semantic information directly.
Markus Krötzsch
12:34:21
What do you need?
liding2006
12:34:22
Jie Bao is the person who work on that
Harold Solbrig
12:34:22
The latter is a bigger challenge, but we'd be willing to help make it work.
liding2006
12:34:32
http://tw.rpi.edu/wiki/Jie_Bao
Denny Vrandečić
12:34:38
harold: you mean to actually store the triples in the back-end, replacing the mysql storage engine?
Daniel Hansch
12:34:43
ok - Thomas Schweitzer is taking care of that on our side.
Markus Krötzsch
12:34:45
Harold: Yes, the latter is a big challenge.
Harold Solbrig
12:34:53
Last I looked, you had a set of tables that looked 98% like a triple store, but it wasn't. I couldn't run Jena on it, for instance.
Markus Krötzsch
12:35:16
It is a triple store
Denny Vrandečić
12:35:24
harold: right. daniel can tell you a bit on connecting it to jena, but indeed, basically just for reading
Markus Krötzsch
12:35:29
but you cannot exchange the DB backend among triple stores ;-)
Harold Solbrig
12:35:32
It's changed since 1.03?
Markus Krötzsch
12:35:36
a lot
Harold Solbrig
12:35:40
Excellent.
Denny Vrandečić
12:35:40
so there is no api for writing triples to smw right now
Markus Krötzsch
12:35:51
oh, but not so much since 1.3, if you mean this
Markus Krötzsch
12:36:17
do never read the DB directly, it will only give you headaches, and it *will* break.
Harold Solbrig
12:36:22
If it is a real triple store (Jena enabled), then writing isn't the problem. (Jena has a write API). The challenge, of course, is making the new info available on the wiki side.
Denny Vrandečić
12:36:36
right
Harold Solbrig
12:36:49
We would like to help with that part - we believe that we can make it work (well)
Markus Krötzsch
12:36:52
Well, it stores and queries triples, it does not supply APIs like a proper store
Markus Krötzsch
12:37:05
I though you meant the internal data processing
Denny Vrandečić
12:37:08
harold: that would be great
Harold Solbrig
12:37:09
Exactly - there is a "triple store" and there is a Triple Store.
Markus Krötzsch
12:37:12
this is really a lot like in a triple store
Harold Solbrig
12:37:35
It is a lot like a triple store - so much, in fact, that it makes no sense that it *isn't* one.
Markus Krötzsch
12:37:44
but direct writing would require to have "another" triple store, since you cannot write back to MediaWiki in any obvious way
Denny Vrandečić
12:37:56
harold, i'm noting down to get back to you in order to discuss this further
Markus Krötzsch
12:38:03
(y)
Harold Solbrig
12:38:05
Fortunately, you've got that wonderful bridge table. Lets discuss offline...
Markus Krötzsch
12:38:27
ok
Anna Levitt
12:38:37
this is Ethan. Who is down to brainstorm some apps?
Harold Solbrig
12:38:37
[Slide - RPI Map Architecture (5)]
baojie_iowa
12:38:48
could somebody add "mtgreaves" in? thanks
Anna Levitt
12:39:07
seems like good apps that generate useful data is an important first step
Denny Vrandečić
12:39:47
mark is in
Harold Solbrig
12:39:57
Featurs - slide 6
Steven Forth
12:39:59
Monitor (LeveragePoint) would be happy to host some brainstorming sessions on a regular basis, and of course we can do this over drinks and tomorrow.
Denny Vrandečić
12:40:01
ethan: not sure i understood what you mean
Mark Greaves
12:40:03
thanks denny!
liding2006
12:40:32
harold:thanks
baojie_iowa
12:40:38
RPI Map: http://map.rpi.edu
Harold Solbrig
12:40:45
Slide 8 - Two Parser Functions
Denny Vrandečić
12:42:21
where are the slides actually?
Denny Vrandečić
12:42:24
the map slides
Harold Solbrig
12:42:36
Another brainstorming topic - I'm concerned about the template syntax. It is becoming increasingly complex. Should we be supporting an XML analog?
Harold Solbrig
12:42:54
Integrated KML file (11)
Harold Solbrig
12:43:10
Questions slide - 12
Marc Feickert
12:43:36
Well, even some api to create a template would be good. Yaron gives one specifically for forms, but a wizard of some sort would be good
Denny Vrandečić
12:44:27
harold: the problem is that the template stuff is mediawiki specific, not smw specific
Harold Solbrig
12:44:36
SMW and Me (slide 2)
Denny Vrandečić
12:44:42
so this would need to include a wider audience
Denny Vrandečić
12:44:53
which slides are on right now?
liding2006
12:45:00
2
Markus Krötzsch
12:45:04
URL?
Harold Solbrig
12:45:08
Jie Bao is the presenter - what is the title and URL?
liding2006
12:45:14
http://tw.rpi.edu/wiki/Image:2008-11-18_wikiBoston-bao.ppt
Steven Forth
12:46:00
A simpler way to support templates and forms and a clearing house to exchange them would be very valuable. But we will need to establish better approaches to documentation.
liding2006
12:46:46
slide 6
liding2006
12:47:01
8
Mark Greaves
12:47:26
Steven, we can host a template directory & clearing house at smwforum.
liding2006
12:47:41
good point
Yaron Koren
12:47:58
There's already one on smw.referata.com - see "Packages'.
Steven Forth
12:48:31
We should begin doing this, the interplay between SMWForum and Referata should be worked out, maybe at this meeting.
Harold Solbrig
12:48:33
Can we put slides on the referata site?
liding2006
12:48:40
slide 12
Harold Solbrig
12:48:42
(.ppt or .pdf format?)
Daniel Hansch
12:48:43
 :-) right
Yaron Koren
12:48:46
Yeah, sure.
Yaron Koren
12:49:02
PDF is better - I don't think PPT is allowed.
Harold Solbrig
12:49:22
This looks similar to SWRL - again, should we be creating a new syntax?
Markus Krötzsch
12:49:31
would like to comment on this work; someone please schedule me for the discussion ...
Markus Krötzsch
12:49:45
(at the end)
Yaron Koren
12:50:08
Do you want to talk?
Denny Vrandečić
12:50:14
do you hear us?
Yaron Koren
12:50:18
No.
Yaron Koren
12:50:36
Okay, we hear you.
Harold Solbrig
12:50:52
Better sound than what we've got locally!
Markus Krötzsch
12:51:23
no more sound hrer
Denny Vrandečić
12:51:29
we dont hear you
liding2006
12:51:30
http://tw.rpi.edu/wiki/Image:MetaModeling_Boston_Nov22-08.pdf
Yaron Koren
12:51:43
How about now?
Denny Vrandečić
12:51:45
we got disconnected
Yaron Koren
12:51:49
Oh.
Denny Vrandečić
12:51:50
please call me again, yaron
liding2006
12:51:56
sorry about that
liding2006
12:52:00
remember our experience in iswc
Markus Krötzsch
12:52:02
feel free to continue the tlk, we will be back
Markus Krötzsch
12:52:17
(sound is back)
Yaron Koren
12:52:23
Cool.
baojie_iowa
12:52:25
http://tw.rpi.edu/wiki/Image:MetaModeling_Boston_Nov22-08.pdf
liding2006
12:52:25
voice skype took much bandwith
liding2006
12:52:41
slide "feature"
liding2006
12:53:13
slide 4
liding2006
12:53:19
slide 5
liding2006
12:53:56
slide 7
liding2006
12:54:41
the last slide indicate the OWL vocabularity which are not supported currently.
Denny Vrandečić
12:54:51
thx
Markus Krötzsch
12:54:56
ok, now I try to have a question
liding2006
12:54:58
http://tw.rpi.edu/dev/cnl
Daniel Hansch
12:55:01
what was the name of the presenter?
baojie_iowa
12:55:09
Zhenning Shangguan
liding2006
12:55:15
Zhenning Shangguan

Tetherless World Constellation, RPI

Daniel Hansch
12:55:16
thx
liding2006
12:56:29
now discussion SMW and OWL2
liding2006
12:57:03
OWL2 DL semantics
Harold Solbrig
12:57:29
I'll upload some slides - it is interesting to see what folks have had to do with OWL semantics when it comes tot he "lexical" section (definition, synonyms, ...). OWL 2.0 helps, but we need SKOS, DC, FOAF, and assoicated roadmap.
liding2006
12:57:46
my understanding, our current stategy is to make SMW as the storage, but not inference engine
Harold Solbrig
12:58:02
There is also the non-OWL community (SKOS and ad-hoc)
liding2006
12:58:13
yes
Steven Forth
12:58:21
We will keep going to about 1:15 and then break for lunch.
Marc Feickert
12:58:41
Ahh, glorious respite!
Markus Krötzsch
12:59:12
(y)
liding2006
12:59:24
harold, we are targetting at meta modeling for other languages as well
Harold Solbrig
13:00:42
http://smw.referata.com/wiki/Image:Importing_Terminology_and_LexWiki.pdf
Harold Solbrig
13:01:01
There is a mention of our preliminary work w/ XMI as well...
Mark Greaves
13:02:27
Another item relevant to the semantics discussion is Ontoprise's work on rules for the Halo Extension. We have a Simple Rules extension designed by Raphael Volz and Benjamin Grosof and others.
Markus Krötzsch
13:03:26
Intersting. We have done research on OWL-Rule-integration, based on prior work by Raphael, so I am confident this will fit with our ideas
Mark Greaves
13:04:08
Have Raphael give you a demo. Daniel Hansh can coordinate. It's pretty cool. They use Ontobroker as the rule execution engine.
Harold Solbrig
13:04:29
I'm glad to see JSON (an XML analog) - again, template syntax isn't (and, imo, shouldn't) be standard.
Guoqian Jiang
13:04:39
is the rule extension for smw?
Denny Vrandečić
13:04:54
true, the template syntax sucks heavily :)
Mark Greaves
13:04:54
Yes
Harold Solbrig
13:05:46
I wouldn't go that far. It is pretty useful, but I don't want to have to create template editing tools, the TSLT (Template Schema Transformation Language), TQuery, etc.
Markus Krötzsch
13:06:10
I get a nice live demo on http://wiki.geekworks.de/mediawiki/index.php?title=USA
Markus Krötzsch
13:06:21
It works well from our location :-)
Denny Vrandečić
13:06:35
harold: the templates should be mostly black boxes
Harold Solbrig
13:06:35
And we are already reaching the level of complexity that writing and debugging templates is non-trivial.
Harold Solbrig
13:06:45
Someone has to write them.
Guoqian Jiang
13:06:45
in our use case, we are interested at rule authoring using smw
Denny Vrandečić
13:06:54
yes :(
Steven Forth
13:07:12
You know, it is actually encouraging to see all of the work from RPI, we should encourage other organizations and students to develop SMW apps, maybe even organize a competition of some kind!
Guoqian Jiang
13:07:15
any extension or use case already there?
Markus Krötzsch
13:07:31
extension is forthcoming
Mark Greaves
13:07:44
Guoqian: see me or Daniel or Henrik at lunch. We have some use cases.
Guoqian Jiang
13:07:52
cool
Markus Krötzsch
13:08:04
sorry, maybe I answered to the wrong question
Marc Feickert
13:08:05
Well, students need to get their degrees somehow... might as well make it useful
Markus Krötzsch
13:08:10
I lost context ;-)
Harold Solbrig
13:08:30
Slide 14 as far as the main presentation
Markus Krötzsch
13:08:54
that I got; thanks
Markus Krötzsch
13:10:38
The JSON export will go into SMW core, so you can draw data from all SMW sits
Markus Krötzsch
13:10:45
(once it is available)
Neil Davies
13:10:53
very nice
Markus Krötzsch
13:13:01
(I have trouble getting this question, is it about vocabulary mapping?)
Markus Krötzsch
13:13:09
(ok, anyway, enjoy your lunch!)
liding2006
13:13:16
break up
Denny Vrandečić
13:13:29
we are not returning after your lunch
Markus Krötzsch
13:13:32
Denny and I will be away for the afternoon. Have a productive meeting!
Denny Vrandečić
13:13:47
it was fun to listen!
Yaron Koren
13:13:51
Thanks for attending!
liding2006
14:11:42
starting of afternoon meeting
Steven Forth
14:12:45
Starting up again
liding2006
14:13:50
presentation on LexWiki
liding2006
14:13:58
by Mayo Clinic
liding2006
14:14:26
Guoqian Jiang , Robert Freimuth, ,Harold Solbrig
Yaron Koren
14:57:45
(Corporate usage)
Harold Solbrig
15:04:24
One of the differentiating characteristics of Wiki vs. GForge, Drupal, etc. is the notion that a Wiki is "organic" all the way down. The data, the forms, the pages and the underlying software all evolve in the same "wiki way". For those who know how to use it, it enables quick change and adoption
Marc Feickert
15:04:53
That's the trick: Helping people know how to use it
Harold Solbrig
15:06:29
Guoqian didn't show it, but we present a dual approach - forms based combined with free wiki editing. It will be interesting to see which one gets more use...
Marc Feickert
15:06:54
We plan to do the same
robertrfreimuth
15:06:58
Marc: The VKC is/will be publishing docs and training materials for non-technical users on the use of LexWiki (and related flavors in use by NCI). Personally, I would like to see more general materials on SMW included in that resource list.
Marc Feickert
15:07:24
ditto
Marc Feickert
15:07:50
We are building video tutorials with Camtasia for many items.
robertrfreimuth
15:07:59
ditto ;-)
Michael Cariaso
15:08:08
I've also been quite happy with camtasia
Marc Feickert
15:08:13
Our user group needs much more than words!!
Marc Feickert
15:08:41
We figured out how to convert camtasia files ot .mov movies, which could allow us to embed them directly into the wiki pages
Harold Solbrig
15:08:47
I worry that the extensive use of templates will reduce the usability of the native wiki. As an analogy, I learned Unix on the original csh. I stepped out for a couple years and when I came back, there was bash. It was powerful, but no one did a simple "ls" any more!
Michael Cariaso
15:08:54
screencasts get out of date very quickly, and updating a movie is a lot tougher than updating video
robertrfreimuth
15:09:03
(again, tool-specific thus far. Government funding rules and the scope of our contract will likely restrict how widely we can go to create the materials, but we will be able to link to/etc any other materials that are out there)
Harold Solbrig
15:09:24
It is quite a challenge to edit Wikipedia these days. It (imo) is no longer approachable by the great unwashed.
Michael Cariaso
15:09:39
harold, how so?
Harold Solbrig
15:09:57
Templates everywhere. Many pages are nothing but templates.
Marc Feickert
15:10:03
Yeah, we need forms and temnplates t capture essential classification information, but we need free text to elaborate and fuel search as well
Mollie Forbes
15:10:14
with a WYSIWYG
Harold Solbrig
15:10:17
If you aren't savvy w/ the template language, it is difficult to know how to proceed.
Marc Feickert
15:10:22
(nod)
robertrfreimuth
15:10:56
Marc: I'll check with those that create our demos. We may be contacting you for more info about the conversion process (if they don't already know about it).
Marc Feickert
15:11:19
sweet. I'll get you in touch with the right guy.
Michael Cariaso
15:11:28
if you click the general top level edit I can agree, but the section specific 'edit' tags (ie http://en.wikipedia.org/wiki/Olpc Mission/History/etc) tend to be fairly template free
Yaron Koren
15:11:41
oceanteacher.net
Michael Cariaso
15:11:45
not obvious enough for new users perhaps
Guoqian Jiang
15:14:03
just like sparql query language, it is powerful but few of people can do that
Marc Feickert
15:14:30
Who's sparkling? 8-)
Marc Feickert
15:15:08
One thing I am considering is using a form to help users create common fixed queries
Michael Cariaso
15:15:21
I use sparql a fair bit, but yes clients immediate say "what do you mean I have to learn a new language?!"
Marc Feickert
15:15:26
lists of counties in a state, records in a repository, etc
Mark Greaves
15:15:34
oceanteacher.net is very slow right now...
Marc Feickert
15:15:47
I am experiencing the same, Mark
Marc Feickert
15:15:57
hey, it painted!
Mark Greaves
15:16:10
it's now up for me.
liding2006
15:16:17
me too
Harold Solbrig
15:16:27
One point that we've made to the NCI is that the expensive part of their ontology creation is the subject matter experts. We want to encourage them to enter as much information as possible and let the "ontologists" worry about formalization after the fact.
Harold Solbrig
15:16:41
W/ feedback of course...
Marc Feickert
15:18:16
That's something I need t talk to you about. Our really big topic is around how much to lock down and how much to leave open in traditiona wiki sense
Mark Greaves
15:18:21
Harold: SMW+ has an implicit division of responsibility between KE's (who have gardening roles in the wiki, and can make global ontology changes), and SMEs (who contribute all the content and schemas)
Guoqian Jiang
15:20:46
is Daniel's slide available somewhere?
Yaron Koren
15:21:35
I'll ask him to upload it.
Harold Solbrig
15:21:57
Yes - Halo evolved concurrently. They saved us much time, but we don't take advantage of nearly as much as we should be.
liding2006
15:22:51
topic: what is path to corperate adoption
Michael Cariaso
15:26:21
sees most folks hadn't heard of perlwikipedia see the code samples at http://www.snpedia.com/index.php?title=SNPedia:FAQ#Where_can_I_get_a_list_of_all_the_snps_in_SNPedia.3F
Yaron Koren
15:33:19
https://grailwiki.grailresearch.com/wiki/index.php
liding2006
15:33:44
login required
Michael Cariaso
15:33:46
requires a login
Neil Davies
15:33:47
it requires login
Yaron Koren
15:34:02
Oh, does it require a login? :)
Neil Davies
15:34:04
yup
Marc Feickert
15:34:10
Yaron...
Marc Feickert
15:34:15
it requires a login
Neil Davies
15:34:18
Steven will demo - its a Monitor wiki
baojie_iowa
16:18:29
Ask to editor, I also tried WikEd, which works well
baojie_iowa
16:19:17
http://en.wikipedia.org/wiki/Wikipedia:WikEd
Steven Forth
16:49:22
Interesting (understandable) choice to rely so much on Ontobroker. A blended open source and proprietary solution.
Guoqian Jiang
16:58:55
http://www.communitywiki.org/cw/DublinCoreForWiki see dublin core for wiki rdf meta data
Guoqian Jiang
16:59:34
it would be usefule to get this rendered into triple store
liding2006
17:31:47
http://smwforum.ontoprise.com/smwforum/
Daniel Hansch
17:31:51
http://smwforum.ontoprise.com/smwforum/index.php/Main_Page
liding2006
17:45:36
best practices
Yaron Koren
18:34:30
"True" n-ary relations: see this page (especially the bottom), and the corresponding talk page:
Yaron Koren
18:34:32
http://semanticweb.org/wiki/N-ary_relations
Daniel Hansch
18:36:41
http://smwforum.ontoprise.com/smwforum/index.php/Study_n-ary_relations
Daniel Hansch
18:36:41
please see also a study for n-ary relations (incl how to display them) in smwforum:
2008-11-23
Michael Cariaso
11:04:19
hello world?
patrick,sbrzesny
11:04:23
hi :-)
Joel Natividad
11:04:30
hello
hsolbrig
11:04:56
Hello Michael
Daniel Hansch
11:05:37
hi - thax for taking me on the list again
baojie_iowa
11:31:27
http://smw.referata.com/wiki/SMW_users_meeting#Features
Daniel Hansch
11:34:59
http://smwforum.ontoprise.com/smwforum/index.php/Triple_Store_Connector
Daniel Hansch
11:43:26
there's abviously a prob with the above article in Firefox
Daniel Hansch
11:43:38
while I am fixing it, pls access the smwforum with IE
Daniel Hansch
11:43:42
apologies!
Daniel Hansch
11:44:45
I fixed it! This article can now accessed with Firefox.
Harold Solbrig
11:45:33
Just like that? (It works in Chrome, btw)
Michael Cariaso
11:45:40
http://www.mediawiki.org/wiki/Extension:Group_Based_Access_Control
Harold Solbrig
11:53:51
Just encountered this on the web and thought I'd pass it on - Semantic MediaWiki got mentioned in the W3C document (my guess is Markus) - go to http://www.w3.org/TR/cooluris/#examples and scroll down.
Joel Natividad
12:07:26
http://www.openldap.org/
Michael Cariaso
12:16:48
I've added 2 notes to the bottom of http://smw.referata.com/wiki/Minutes_for_November_22_2008_SMW_Users_Meeting_at_Monitor_Group
Michael Cariaso
12:17:21
one, points out a bug I've observed at http://smw.referata.com/wiki/Special:Categories
Michael Cariaso
12:17:53
the other is a concrete example of a bot which reads&writes to smw.referata.com
Michael Cariaso
12:17:55
http://smw.referata.com/wiki/Bots
Daniel Hansch
12:26:08
pls see this article for a halo-extension feature which allows a "watch" on semantic data:
Daniel Hansch
12:26:09
http://smwforum.ontoprise.com/smwforum/index.php/Semantic_Notifications
Michael Cariaso
12:27:21
workflows: http://pipes.yahoo.com/pipes/pipe.edit?_id=avkEShi32xG_EF6KZVUMqA is the interfaces uses work with (very javascripty) to create 'workflows' of a sort
Harold Solbrig
12:27:51
Chat room?
Harold Solbrig
12:28:20
Oh - just took a while to get through. Sorry
Daniel Hansch
12:29:04
hi mike
Michael Cariaso
12:33:02
http://smw.referata.com/wiki/Workflows
baojie_iowa
12:44:21
http://smw.referata.com/wiki/SMW_users_meeting#Features has the summary
Daniel Hansch
13:42:42
Feel free to use the discussion board here for communication:
Daniel Hansch
13:42:43
http://smwforum.ontoprise.com/smwboard/index.php
Michael Cariaso
13:54:52
http://www.gmane.org/ archives mailing lists, provides good search, RSS, and even exposes an NNTP interface for folks who remeber that. http://dir.gmane.org/gmane.org.w3c.semantic-web is a sample discussion group
Daniel Hansch
14:00:06
harold: I need your contact details :-)
Harold Solbrig
14:00:17
solbrig.harold@mayo.edu
Harold Solbrig
14:00:30
If you need more, tell me where you want it filled in
Daniel Hansch
14:01:17
thats fine. thank you very much!
Retrieved from "http://tw.rpi.edu/wiki/Boston_SMW_users_meeting"
Personal tools