deegree PSC meeting, 2010-05-06, IRC #deegree
(18:56:12) JensFitzke: Hi all, the agenda for today is here:
(18:56:13) JensFitzke: https://wiki.deegree.org/deegreeWiki/PscMeeting/PscMeeting20100506
(18:56:26) JensFitzke: We start as soon as Klaus is here
(19:00:43) kgreve [~4fe0e8d6@gateway/web/freenode/x-nyepidkdropbckad] hat den Raum betreten.
(19:01:25) kgreve: anybody on board?
(19:01:43) ampoth: yes
(19:01:45) JensFitzke: we're complete and I'd say ready to start
(19:02:02) JensFitzke: Welcome to the PSC meeting
(19:02:12) JensFitzke: The agenda is here https://wiki.deegree.org/deegreeWiki/PscMeeting/PscMeeting20100506
(19:02:29) JensFitzke: Anything else to be added by now?
(19:03:03) kgreve: not from my side
(19:03:13) JensFitzke: If this is not the case for the moment, we could start with (805) OGC WFS compliance
(19:03:16) Herman_: I send an email a few minutes ago: How to make more people familair with deegree?
(19:03:23) JensFitzke: I guess this is mine, so ....
(19:04:11) JensFitzke: ... we found a bug in the respective test suite and we're currently trying to get this fixed by the OGC in order to be able to pass the compliance test. Please don't ask for details.
(19:05:00) JensFitzke: Good news btw is that we (lat/lon) have an agreement now with the OGC to support them within the CITE work. There'll be a press announcement about that quite soon.
(19:05:13) JensFitzke: Any questions or concerns?
(19:05:21) ampoth: no
(19:05:24) Herman_: no
(19:05:25) kgreve: no
(19:05:58) JensFitzke: Next agenda item is (838) Release management guidelines
(19:06:24) JensFitzke: There we had a decision in our last meeting: " IDgis to review German deegree release guidelines and translate into English. To be supported by lat/lon. " -- any progress there?
(19:07:18) ampoth: I sent guidelines to Herman some weeks ago
(19:07:23) Herman_: Our employee, Arie Kraak made start with it, but did not yet complete it because of other priorities (preparing the exhibition of last).
(19:07:48) JensFitzke: Can you give an idea when this could be completed?
(19:07:59) Herman_: He will finish this work next week, currently he is on holiday
(19:08:25) kgreve: sounds good
(19:08:28) JensFitzke: Sounds good. So we'll have that available for the next meeting.
(19:08:44) Herman_: Yes, it will be available then
(19:08:50) JensFitzke: Next agenda item is (1126) new ECW libraries for deegree
(19:09:16) JensFitzke: There we had a decision: "Herman to prepare a report on JPEG2000 implementation issues for next PSC meeting and to update the tracker item accordingly. " -- any progress?
(19:10:14) Herman_: Sorry, to be honest I forgot more or less. I will discuss this with Reijer Copier to be sure it is available next meeting.
(19:10:22) JensFitzke: OK
(19:10:37) JensFitzke: Next agenda item is (1253) Finalize process documentation
(19:10:44) JensFitzke: Respective decision was: "Andreas, Herman, and Klaus to review the documents listed in the tracker item. PSC to consider the documents in the next meeting." -- any progress?
(19:12:01) kgreve: I tried to find out id other OSGeo Project have similar documents, did´nt finde much
(19:12:55) kgreve: think we have to complete the documents adopted from Mapsserver on our own
(19:13:01) ampoth: I had a short look at it and did not relly understand what to do and after this - I must confess - I forgot to ask
(19:13:59) JensFitzke: Any proposal how to proceed?
(19:14:28) Herman_: I jusr read deegreeCommitterGuidelines. I still have to read others. Some hints how to evaluate these documents are welcome.
(19:15:23) kgreve: could Judith or some else do a small ammount of research and find some materiale we can adopte ?
(19:15:36) JensFitzke: hmm, they should be written in a way that they give understandable guidelines to newbies, they shall be as comprehensive as possible, they should not contradict each other
(19:15:38) Herman_: Does anyone have links to comparable documents in other open source projects?
(19:16:23) JensFitzke: These documents are already copied (quite some contents) from other projects. This is stated in the documents.
(19:16:59) kgreve: sure - but we need more
(19:17:07) JensFitzke: I do not think that we need more copy/paste. IMHO it just needs some thorough reading, thinking and some adjustments. Doesn't it?
(19:17:16) JensFitzke: @kgreve: more of what?
(19:17:31) kgreve: material
(19:17:36) kgreve: formulations
(19:18:27) kgreve: I think we need some kind of brain storming - what do we really need, what do we want?
(19:19:00) JensFitzke: Sounds like we need a workshop meeting then. But that's quite a hurdle.
(19:19:12) kgreve: If we have some more or less professionel examples, this will be more simple than the other way
(19:20:17) kgreve: May be we can circulate some material and drafts
(19:20:40) Herman_: How about accepting the current documents and change as needed, depending on experience?
(19:20:51) JensFitzke: I'll see if one of our team members could do some work on the documents.
(19:21:14) JensFitzke: @Herman: There are some points which need to be finalized. They cannot be left like they currently are.
(19:21:18) kgreve: I will support
(19:21:43) JensFitzke: Any more questions or concerns regarding this issue?
(19:22:06) kgreve: I think the points are as they are because we have no opinion
(19:22:13) Herman_: OK, let's finish the open points. Can aynone help to point me to those points?
(19:22:24) JensFitzke: I can
(19:22:29) JensFitzke: ... will do
(19:22:49) JensFitzke: ... hold on ...
(19:23:23) JensFitzke: ... look at everything which is bold and italics on https://wiki.deegree.org/deegreeWiki/StructuresAndProcedures
(19:24:35) Herman_: OK, should we edit the wiki page or send changes by e-mail?
(19:24:48) JensFitzke: Pls edit the page directly
(19:24:53) Herman_: OK
(19:25:17) JensFitzke: Looking forward to the results then. Can we move to (1297) OSGeo incubation follow-up ?
(19:25:27) kgreve: yes
(19:25:31) Herman_: yes
(19:25:58) ampoth: ok
(19:25:59) JensFitzke: Decision was: "Regular reports to OSGeo to be prepared in rotating responsibility. Andreas to prepare the first report for the next PSC meeting."
(19:26:17) JensFitzke: Andreas and me discussed that one. After rethinking I came to the conclusion that obviously the annual reports were meant. There are no shorter reporting periods. I'd suggest that I prepare a draft annual report and hand it over to you for review before the next meeting. There wasn't a reminder yet from OSGeo ...
(19:26:53) Herman_: OK
(19:27:01) kgreve: o.k. sounds realistic
(19:27:21) ampoth: ok
(19:27:47) JensFitzke: Next one is then (1304) Review organizational structure
(19:28:00) JensFitzke: There, the tracker says: "The PSC shall review the project's organizational structure and discuss product management. The procedure for OGC compliance
(19:28:00) JensFitzke: certificate application may be taken as an entry point for the discussion."
(19:28:42) JensFitzke: I must confess that I did not make any progress on that in the meantime... did you?
(19:28:58) kgreve: no
(19:29:50) Herman_: No, are there specific questions/problems regarding organizaional structure?
(19:30:42) JensFitzke: Not really. You could regard it as an issue to open up the project even more. But perhaps that's a bit too much for the moment...
(19:31:23) JensFitzke: That came from the OGC compliance fees where they consider the annual turnover of the company applying for the compliance certificate.
(19:31:38) JensFitzke: ... but for an open source project there could be arguments...
(19:32:09) JensFitzke: If there are no strong opinions currently I'd say we simply postpone it for a while and reconsider as we feel the need
(19:32:44) Herman_: Perhaps we should think of specific actions to involve more people in development?
(19:33:23) JensFitzke: Would that fit into your new agenda item?
(19:33:25) kgreve: sounds like a plan (?)
(19:33:51) Herman_: Yes, it is related to the new aganda item
(19:34:24) JensFitzke: Ok, then... we have Christian's (1470) Change Request: Change URL pattern of servlet mapping from 'services' to 'deegree'
(19:34:34) JensFitzke: ... where the tracker says: "Summary of change: change servlet mapping for the URL pattern from 'services' to 'deegree'. "
(19:34:58) Herman_: Good idea, I agree.
(19:35:00) kgreve: good idea, any problems with that?
(19:35:07) ampoth: I disagree because first part of the path names the context and second the application. In my opion the context is deegree and the application is WFS, WMS etc. or services if more than one deegree service is configured within one application
(19:35:57) JensFitzke: Wasn't just a proposal for the default configuration?
(19:36:21) JensFitzke: Wasn't that (Christian's issue) just a proposal for the default configuration?
(19:36:30) ampoth: sorry I do not understand the question
(19:36:36) JensFitzke: ... which could be changed by any user?
(19:36:51) JensFitzke: ... or is that kind of hard-coded?
(19:37:48) Herman_: No, it's a default configuration which could be changed by any app administrator.
(19:38:16) ampoth: I think what is intended by Christian ist configuration of demos
(19:38:40) JensFitzke: If Herman is right we would just increase name-dropping, which in general is a good idea.
(19:39:27) kgreve: I agree with Jens
(19:39:43) ampoth: how do we increae 'name dropping' if we follow Christians suggestion?
(19:40:08) JensFitzke: @ampoth: Christian is referring to d3: "This change request should apply for deegree3"
(19:40:29) ampoth: I still do not understand
(19:40:47) Herman_: For example, a url would then be: http://www.lat-lon.de/mywebapp/deegree?request=GetCapabilities&service=WMS
(19:40:51) JensFitzke: @ampoth: don't understand what?
(19:41:16) ampoth: the benefit
(19:41:26) JensFitzke: name-dropping
(19:41:42) ampoth: can you give an example?
(19:41:51) JensFitzke: Herman did
(19:42:14) kgreve: see the word deegree where a service is provided by a deegree installation
(19:42:48) JensFitzke: question is obviously if context name or servlet mapping shall be deegree by default ... uff
(19:42:57) ampoth: what is the benefit of http://www.lat-lon.de/mywebapp/deegree?request=GetCapabilities&service=WMS compared to http://www.lat-lon.de/deegree/wms?request=GetCapabilities&service=WMS
(19:43:24) JensFitzke: @ampoth: what if the "wms" servlet acts as a WFS as well?
(19:43:50) ampoth: http://www.lat-lon.de/deegree/services?request=GetCapabilities&service=WMS
(19:44:17) JensFitzke: ... and the proposal was to change the servlet mapping from "services" to "deegree"
(19:44:34) JensFitzke: ... and I do not know what shall happen with the context name...
(19:44:41) JensFitzke: Do we need more information here?
(19:44:51) Herman_: Mostly you will have multiple web-applications using deegree, so you can't (or won't) use deegree as the name of the webapp.
(19:44:57) ampoth: this also would allow to have:
(19:44:58) ampoth: http://www.lat-lon.de/deegree/wms?request=GetCapabilities&service=WMS
(19:45:00) ampoth: http://www.lat-lon.de/deegree/portal?request=GetCapabilities&service=WMS
(19:45:27) ampoth: sorry last URL must be http://www.lat-lon.de/deegree/portal
(19:45:49) JensFitzke: I have the feeling that we need more information here. At least for me that's true. I thought I understood the whole impact but in fact I do not.
(19:46:33) JensFitzke: It's just about names, but as always names are important ...
(19:46:54) Herman_: OK, let's ask some developer for pros and cons
(19:46:56) JensFitzke: If you agree, I'll ask for some explanation about the intention and impact
(19:47:01) kgreve: o.k., so we look for more information and postpone to next meeting
(19:47:10) ampoth: ok
(19:47:15) JensFitzke: ... and the possible changes a user could apply after the default installtion.
(19:47:24) Herman_: ok
(19:47:36) JensFitzke: poooh, names...
(19:47:44) JensFitzke: OK. Next one is (1474) FOSS4G WMS shootout participation
(19:48:07) JensFitzke: That's a hard one I guess. Let's nevertheless try to make it short.
(19:48:31) JensFitzke: Did everybody read the respective mail(s)? Are there any questions?
(19:48:36) kgreve: am I right if I think it is quite a simple benchmark?
(19:48:47) ampoth: yes - no
(19:48:53) kgreve: ???
(19:49:12) JensFitzke: @kgreve: I do not have any detailed information about the shootout, sorry.
(19:49:33) JensFitzke: ... it's just the label I do not like, "shootout'"
(19:50:03) Herman_: Yets, I read the shootout results from last year. Do deegree developers expect that deegree3 can achieve performance results comparable to geo-server and mapserver?
(19:50:04) ampoth: @kgreve: yes to Jens' first question, no tho his second
(19:50:25) kgreve: but it´s about speed or rsponce time?
(19:50:34) kgreve: responce
(19:51:27) Herman_: Last year the measure was throughput, number requests per second, under same conditions for all wms's
(19:53:13) kgreve: I would expect a simple WMS is better than a complex one, if you measure throughput
(19:54:15) Herman_: Is the character of the shootout of this year already known?
(19:54:36) JensFitzke: @Herman: As far as I know there won't be a change in the procedure
(19:54:51) JensFitzke: ... but there is a plan to change the procedure for the future...
(19:55:00) JensFitzke: ... if I am right.
(19:55:45) kgreve: My problem with this thing is: Speed or throughput is important but not our main goal, wy should we compete in this field
(19:55:48) JensFitzke: TMC told me that d3 WMS is not in the position to really compeed with the other products. They are (supposed to be) profiled, deegree is not (yet).
(19:56:31) JensFitzke: upps, shall be "compete"
(19:56:35) kgreve: So: Not our cup of coffee
(19:56:36) ampoth: I do not have a clear option on this. In general if have doubts on usefullness of such benchmarks because it's like comparing äpfel mit birnen (sorry acn't say this in english). But not joining the shootout may looks like we do not trust our own software.
(19:57:34) JensFitzke: My problem is that in the end, only the simple stories make it to the general public.
(19:57:40) Herman_: Can d3 developers do the benchmark of last year to see of d3 is comparable to other wms's?
(19:58:06) kgreve: I think it is not too complicate to communicate that this kind of benchmarking is too simple
(19:58:16) JensFitzke: They could kind-of participate in the shootout without being mentioned, I guess.
(19:58:35) JensFitzke: @kgreve: Yes, but in the end, only the simple stories make it to the general public.
(19:58:47) JensFitzke: ... and only the big ones. No one would like to read our version.
(19:59:29) kgreve: o.k. so Arc Server or Mapserver is #1, what does it matter idf deegree do´nt participate
(19:59:52) kgreve: the simple story is xxx is #1
(20:00:18) JensFitzke: @kgreve: I agree. The simple story wouldn't be: "deegree is not participating in the shootout"
(20:00:40) Herman_: If deegree is not in the results, no one knows that deegree is a competing product.
(20:01:03) JensFitzke: @Herman: This is basically Judit's issue.
(20:01:25) JensFitzke: ... she had something like "deegree could be considered to be dead"
(20:01:45) kgreve: I think we need a story for those how are not satisfied with the simple story
(20:01:58) JensFitzke: @kgreve: think so too
(20:02:13) JensFitzke: @kgreve: I think so too
(20:02:56) JensFitzke: @Herman: Problem with bad results is that no one would consider the explanation.
(20:03:00) Herman_: How much time is there for preparation? Is there a change to do some cirtical optimization before results become public available?
(20:03:33) JensFitzke: There is time, but I think there are no resources. There are much more important things to achieve. API consolidation e.g.
(20:03:53) kgreve: ij every case, and esp. if code optimation is possible, other we be better than us
(20:04:15) kgreve: will be better..
(20:05:08) Herman_: I suppose deegree will be #1 on a feature comparison list?
(20:05:26) JensFitzke: I guess so
(20:06:12) Herman_: Any idea how to tell the feature comparison story?
(20:06:49) JensFitzke: If I understood everything right that's something for the near future. But I do not know about any details.
(20:08:16) Herman_: OK, let's discuss a more realistic comparison of wms's with FOSS4G organization.
(20:08:54) kgreve: so the more than simple story is: degree is a very complex think (see feature list) and can´t be compared with simple WMS. Therefor we do`nt what to partizipate in a simple benchmark
(20:09:05) JensFitzke: Would you like to take a first step here, Herman?
(20:09:22) kgreve: but will support more complex and realistic tests
(20:10:31) kgreve: motion: let´s write an open letter to the shoutouts
(20:11:58) JensFitzke: Who would like to make a start here? Open letter, mail to discuss@osgeo or anything comparable?
(20:12:16) Herman_: @Andreas: What do you think?
(20:14:35) ampoth: sorry it will take a bit to formulate an answer ...
(20:15:27) JensFitzke: be right back
(20:18:49) ampoth: sounds a bit stupid, but I really don't know. In my opinion comparsion between OWS of different vendors will never be meaningful no matter if you have a simple benachmark or a comparsion of a feature list. If for example deegree or UMN is best depends on requirements of a concrete project and not on an more or less abstract list of things that can be done in an benchmark or a general...
(20:18:51) ampoth: ...comparsion of features ... (to be continued)
(20:20:25) JensFitzke: back again
(20:20:26) ampoth: So I do not think starting a discussion at osgeo (or somewhere else) is really usefull because the general problem remains - it just will be moved to another level ...
(20:21:26) JensFitzke: So we could agree that this kind of comparison is not really useful for real-world applications, right?
(20:21:41) ampoth: yes
(20:21:54) JensFitzke: ... so we could write that open letter, right?
(20:22:14) ampoth: I think so
(20:22:47) JensFitzke: Anybody would like to make a start?
(20:22:56) ampoth: .. but I also think it won't change much :-(
(20:23:27) JensFitzke: right, this is a sad world. Lisa Lisa.
(20:23:46) JensFitzke: ... Cat Stevens, as far as I remember.
(20:24:03) ampoth: be right back in a minute
(20:25:33) JensFitzke: As nobody would like to make I start, I'll do so. But I cannot promise, when I'll manage to.
(20:25:59) kgreve: I will support
(20:26:35) JensFitzke: I am counting on the support of all of you. Will send a draft version of the "letter" to the list.
(20:26:38) Herman_: I will help if Jens makes a start
(20:26:46) JensFitzke: Next agenda item is (1494) Release for deegree day
(20:27:01) ampoth: back again
(20:27:19) JensFitzke: ... as far as I understand TMC is already taking a 3.0 release into account. Any questions or concerns?
(20:28:10) kgreve: IDid´nt they promise?
(20:29:00) JensFitzke: We could regard this both as a TMC decision and as a promise.
(20:30:13) kgreve: o.k.
(20:30:13) JensFitzke: [btw. sorry for rushing through the agenda. It's quite some items...]
(20:30:23) Herman_: What will this release include?
(20:30:51) JensFitzke: @Herman, this could be in today's TMC's mail (which I did not manage to read completely)
(20:30:59) JensFitzke: who did?
(20:31:14) ampoth: For it is important that even the first alpha release should be easy to install and configure. Otherwise there is a danger that deegree 3 will have the label: 'difficult and complicated' even if this will be changed in near future
(20:32:31) JensFitzke: @ampoth: I fully agree.
(20:32:56) JensFitzke: Shall I ask the TMC to provide a plan for the 3.0 deegree day release?
(20:33:25) kgreve: yes, please
(20:33:35) Herman_: Would be nice if there is a release even if it only includes minimal, most important functions ,..
(20:34:08) ampoth: I think this is a good idea. meanwhile I will go on with Markus designing a concept for d3 installation and configuration
(20:34:09) Herman_: However, what is released must have good quality.
(20:34:25) Herman_: ok
(20:34:36) JensFitzke: Ok, let's close that one with a request to the TMC for a feature list. And weÄ
(20:34:41) kgreve: yes
(20:35:00) JensFitzke: [sorry] And we'll stress that quality of features is more important than number
(20:36:00) JensFitzke: I'd suggest to move the potential discussion on (1495) What is "stable" in a release? to the next meeting. We're running out of time.
(20:36:02) Herman_: ok
(20:36:12) kgreve: o.k.
(20:36:37) ampoth: ok
(20:36:50) JensFitzke: @all: those OKs where meant for the release thing, right?
(20:37:22) kgreve: right
(20:37:41) Herman_: yes, and also for your suggestion to finish the meeting now
(20:38:00) ampoth: OK
(20:38:01) JensFitzke: @Herman: I did not suggest this.
(20:38:20) kgreve: but it´s a good idea
(20:39:10) JensFitzke: looking at the other business items I'd say there is only Herman's new one to consider...
(20:40:11) JensFitzke: Herman?
(20:40:42) Herman_: My item is added because of my experiences last week during the exhibition: Only a few people ever heard of deegree while al know about Mapserver and Geoserver
(20:40:54) Herman_: al = all
(20:41:05) JensFitzke: Shall we consider this for discussion now?
(20:41:28) Herman_: No, we can postpone it to next meeting
(20:42:39) JensFitzke: Any other business (apart from postponed agenda items)?
(20:43:13) ampoth: is it possible to have telcos instead of IRCs?
(20:43:26) kgreve: not from my side
(20:44:14) JensFitzke: I'd prefer this IRC stuff over telcos as we have a nice protocol withouth doing anything else.
(20:44:31) kgreve: me too
(20:44:42) Herman_: Once a year I can come to Bonn to have a PSC meeting. Would that help?
(20:44:56) kgreve: perfect
(20:44:57) JensFitzke: In those telcos most of the contents tends to vanish soon after the meeting.
(20:45:09) JensFitzke: @Herman: that would be great. When next?
(20:45:32) ampoth: looks like I am the minority :-)
(20:46:12) JensFitzke: OK, I'll prepare the meeting notes then and update the tracker if applicable.
(20:46:31) JensFitzke: And will call for the next meeting sooner than this time.
(20:46:35) Herman_: I like to combine it with a visit to Limburg. I can propose a date as soon as I have a suitable appointment in Limburg
(20:46:57) JensFitzke: @Herman: great.
(20:47:04) JensFitzke: For the next meeting: Any restrictions early June?
(20:47:23) Herman_: early June is ok with me
(20:47:33) kgreve: for me too
(20:48:07) ampoth: can not say because there some possible appointments in first half of june
(20:48:26) JensFitzke: Will make a doodle query for the first week of June then.
(20:48:39) kgreve: o.k.
(20:48:40) JensFitzke: Ready to close the meeting?
(20:48:44) Herman_: ok
(20:48:46) ampoth: ok
(20:48:52) kgreve: ok
(20:49:00) JensFitzke: Wel thanks to all. And again: Sorry for the rush.
(20:49:06) ampoth: bye
(20:49:12) Herman_: bye
(20:49:13) kgreve: Thax, bye