Hi! Sebastian, you nicely addressed most of my concerns and seem to understand the problem.
Tom & Sam: In this thread on Nov 17 I wrote "Many people start problem solving by a using a search engine, [...] the discussions need to be publicly available for reading without login". That seems to be the point you missed but Sebastian understood. Thanks Sebastian! If you already know where to look then you have already done half the job, I tried to address the problem of not knowing where to look (e.g. questions like "is there any openEHR work regarding deafness", such search results should also include content from CKM discussions). Tom you are probably one of the few people that actually know exactly where to look for any thinkable openEHR thing, but you are an exception (in a positive sense) and will seldom be the basis of a standard use-case ;-) Tom wrote: > We have robots turned off for all SVN repositories, including the one that > used to > hold all the archetypes. Once upon a time not even the specifications were searchable since they only were in the indexing-blocked SVN, that was really bad, but it was easily solved by copying releases to a indexable web server directory. Koray: Regarding Ocean as a "social service organisation", closed doors etc, I agree that would be valuable to discuss, but the discussion is neither technical or clinical so I'll try to stop abusing these lists for that purpose and have started a wiki page for the topic at: http://www.openehr.org/wiki/display/oecom/openEHR+transparency (Those that want that discussion in the inbox can subscribe to changes on that page. I'll post something next week but others might post something before me.) Best regards, Erik Sundvall erik.sundvall at liu.se http://www.imt.liu.se/~erisu/ Tel: +46-13-286733 (Mail & tel. recently changed, so please update your contact lists.) P.s. Sebastian: Regarding archetype reviews, yes they would be better presented together with the context they were written in, so I'd suggest that such a view would showing them in context could be created and being the one exposed to search engines, there is no point in publishing every comment as an separate page. Tom and other programmers: Yes, it would be nice to republish at least the completed releases from sourcecode repositories too (in a indexable way) if the SVN server can't take the load directly. Try searching for "openehr" on http://www.google.com/codesearch - it will mostly show results not hosted by openehr.org (and some old defunct now blocked links from openehr.org). It even searches and nicely displays contents of zip-files with code, it identifies programming language, licence type etc. A wonderful tool that openehr.org currently has opted out from. (It even found cached code from your old now taken-over domain Deepthought. That's an example archiving service as I mentioned earlier as a positive thing with being indexable.) On Fri, Dec 4, 2009 at 07:52, Sebastian Garde <sebastian.garde at oceaninformatics.com> wrote: > Hi, > > I think a couple of things could be changed to make CKM more > search-friendly, but I also think we have to look carefully what we want and > where: > > The idea behind linking comments/discussions directly to an archetype and > not just using a mailing list or the like is that we want comments for an > archetype gathered in one place so that they are in fact easier accessible > than via common search engines. > I certainly felt that a lot of discussions were repeated on the mailing > lists, because nobody could remember that the question had been asked for > that archetype already. > However, currently it is only possible to search comments for a certain > archetype and we may want to search for comments of all archetypes (because > we know what we are searching for, but not exactly for which archetype it > may have been posted). > All new comments are already available by a news feed you just need to > subscribe to as well as on the Dashboard. Also you can select to get email > notifications on any new comment (or comments for a specific archetype or > comment for a specific thread of comments.) > There are probably tools to take a newsfeed and post the contents somewhere > for anybody who likes to do this. > However, there may be additional value to automatically post the comments to > somewhere where they are indexable from google etc. This e.g. could be > another openEHR mailing list that receives all new comments. > We could also make the comments available without having to log in to CKM if > people feel strongly about this. > Then only if you want to post a new comment you need to log in and probably > only access the commenters' profiles once logged in as well. > With regard to archetype reviews, I am not so sure if they should be > accessible without even logging in. > Neither I am sure if there is value in having them indexed by search > engines. I strongly believe that they need to be structured and displayed in > an archetype-specific way and groupable by review rounds, directly linked to > the way the archetype looked liked at that stage of the review process, etc. > to be useful. Everybody can access all reviews in CKM (when logged in), > while reviewers can choose to be anonymous (and only reveal their identity > to other members of the review team). Certainly the CKM approach is more > open than any HL7 (or the like) process I have seen. Everybody can access it > and everybody who wants to can participate. > With regards to archetypes, you can download all of them or a selection of > them. There are webservices to access them, and a couple more are going to > be added in the next release as discussed on the wiki and wave. And if there > is need for more, we can always add them > When google wave is out of beta and open to everybody, we will certainly > explore how we can make use of it, integrated in CKM and/or as a starting > point for archetype development, etc. > > Regards > Sebastian > > Thomas Beale wrote: > > I am not sure what use having search engines seeing into the CKM is. We have > robots turned off for all SVN repositories, including the one that used to > hold all the archetypes. Search engines only tell you about things you did > not already know about; whether they could report anything coherent from CKM > I am not sure, same as for all the source code in the SVN repositories. Or > are you suggesting we make all that searchable as well (it kills performance > by the way). > > Google or any other search engine doesn't know how to search CKM in an > intelligent fashion....but if you go into CKM, which is fully open, you can > see everything. I am unclear on the problem. > > - thomas beale > > > Erik Sundvall wrote: > > Hi Sam! > > On Wed, Dec 2, 2009 at 18:20, Sam Heard <sam.heard at oceaninformatics.com> > wrote: > > > Hi Erik > Can you tell me what search capabilities you want in CKM that are not there. > You can export a prot?g? ontology, all the archetypes and have all the > search power we have thought of from the asset management platform. > Unsearchable seems a little unfair. > > > If you read my CKM-search-reasoning in other related messages > carefully again you will see that I have been talking about the CKM > not being searchable via major search engines (like Google Search) > hence the wording "open public searchable space". > > The problem is that the CKM content currently is "locked in" behind > passwords and a search-engine-unfriendly application structure so that > the content is not a proper part of the "web" that search engine > spiders can index. This is a fairly simple technical publishing > problem that can be solved if there is a will from the ones owning the > CKM application. A more serious meta-problem is if the > search-engine-unfriendliness is not seen as a problem by the > application owners and by the foundation using the application. > > The wiki and mailinglists (via archives) do not suffer from this > searchability problem, they were built with openness in mind and are > fully searchable and any discussion regarding certain archetypes in > them etc will be found. An extra plus is that their content can also > be archived by sites like http://www.archive.org > > In openEHR we are often talking about the value of capturing clinical > context in order to interpret data, as a thought-experiment try to > apply the same thinking regarding archetype development. You (and your > search queries) might want to see the context of discussion and the > review comments for archetypes, not just the final archetypes. > > Best regards, > Erik Sundvall > erik.sundvall at liu.se http://www.imt.liu.se/~erisu/ Tel: +46-13-286733 > (Mail & tel. recently changed, so please update your contact lists.) > > > > > -----Original Message----- > From: openehr-technical-bounces at chime.ucl.ac.uk [mailto:openehr- > technical-bounces at chime.ucl.ac.uk] On Behalf Of Erik Sundvall > Sent: 19 November 2009 09:35 > To: For openEHR clinical discussions > Cc: For openEHR technical discussions > Subject: Re: openEHR community on Google Wave > > Hi! > > On Thu, Nov 19, 2009 at 06:48, Heather Leslie > <heather.leslie at oceaninformatics.com> wrote: > > > If I have caused any confusion, I apologise. I'm just enthusiastic > > > and > > > interested to further explore the potential (or not) offered by > > > Google > > > Wave. > > > It is a very nice initiative Heather and there is no need to > apologise, just a need to get the discussions out in open public > searchable space (and that also goes for the currently unsearchable > CKM). > > I believe that in a set of properly managed wave conversations it > might be easier to follow the discussion flow, and it might be a less > fragmented user experience than the current CKM is. If done right and > when there are more wave providers than Google (since wave uses a > truly open protocol) then we could at the same time get rid of the > current CKM vendor lock-in and extension limitations (without creating > another vendor lock in). > > > > While these initial 'coordinating waves' are public, small groups may > > > go off > > > and use a private Wave to work on a task or project - just like they > > > do now > > > using email, skype or IM. > > > Yes of course some conversations (or parts of conversations) will > always be private since humans prefer to work that way sometimes. The > problem is if things are inaccessible and unsearchable even when there > is no intention to keep the discussion private. > > > > The result should be identical - submitting the > draft archetype to CKM or contributing to the email lists or wiki. > > > If wave-based tools become widespread and powerful enough to do > openEHR review, voting etc., then I don't see CKM as a necessary step > in the pipeline to finally submitting archetypes/templates to simple > stable repositories. Every shift of tools along the way adds a > potential user confusion. > > By the way, have you tried using mindmapping gadgets for openEHR > related development in wave, I found an open source mindmapping gadget > that even includes a voting mechanism and freemind-import facilities > at: > http://wave-samples-gallery.appspot.com/about_app?app_id=64007 > See also: http://www.brucecooper.net/2009/11/mind-map-gadget-for- > google-wave.html > And since the mindmapping gadget is open source it could easily be > modified by any java/GWT developer to add features that you'd find > useful for openEHR related use :-) > > Best regards, > Erik Sundvall > erik.sundvall at liu.se http://www.imt.liu.se/~erisu/ ?Tel: +46-13-286733 > (Mail & tel. recently changed, so please update your contact lists.) > > P.s. To add voting to suitable items (e.g. corresponding to when you > use voting in CKM) it seems like > http://wave-samples-gallery.appspot.com/about_app?app_id=23006 might > be useful. I guess a proper discussion will often solve things without > the need for voting though... > _______________________________________________ > openEHR-technical mailing list > openEHR-technical at openehr.org > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical > > > _______________________________________________ > openEHR-technical mailing list > openEHR-technical at openehr.org > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical > >
