Getting it done at LocationTech
This is a modal window.
Das Video konnte nicht geladen werden, da entweder ein Server- oder Netzwerkfehler auftrat oder das Format nicht unterstützt wird.
Formale Metadaten
| Titel |
| |
| Serientitel | ||
| Teil | 135 | |
| Anzahl der Teile | 193 | |
| Autor | ||
| Lizenz | CC-Namensnennung 3.0 Deutschland: Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen. | |
| Identifikatoren | 10.5446/20427 (DOI) | |
| Herausgeber | ||
| Erscheinungsjahr | ||
| Sprache |
Inhaltliche Metadaten
| Fachgebiet | ||
| Genre | ||
| Abstract |
| |
| Schlagwörter |
00:00
Algebraisches ModellLokales MinimumMultiplikationsoperatorURLSoftwareentwicklungSoftwareentwicklerGüte der AnpassungBitGradientVorlesung/KonferenzBesprechung/Interview
00:45
VarianzDienst <Informatik>SoftwareentwicklerRechenwerkWiederkehrender ZustandKurvenanpassungDreiKontextbezogenes SystemOpen SourceGruppenkeimEindringerkennungICC-GruppeRed HatGoogolOrakel <Informatik>Vorzeichen <Mathematik>KraftElementargeometriePatch <Software>CodeProjektive EbeneSondierungKurvenanpassungSelbst organisierendes SystemGoogolZahlenbereichSoftwareentwicklerDeskriptive StatistikQuellcodeURLSummengleichungOrdnung <Mathematik>Vorzeichen <Mathematik>GrundraumProdukt <Mathematik>GrenzschichtablösungProgrammbibliothekElement <Gruppentheorie>Coxeter-GruppeMereologieCASE <Informatik>PerspektiveMetadatenForcingSoftwareOpen SourceRechenschieberKontextbezogenes SystemArithmetische Folgesinc-FunktionGruppenoperationShape <Informatik>ElementargeometrieEinbettung <Mathematik>PunktwolkeGüte der AnpassungInternet der DingeDifferenteVererbungshierarchieRaumfüllende KurveAbgeschlossene MengeCodeComputeranimation
06:23
HilfesystemE-MailBildschirmmaskeMathematikMigration <Informatik>FokalpunktRepository <Informatik>ElementargeometrieWebDAVKraftOpen SourceTreiber <Programm>DatenbankVersionsverwaltungEreignishorizontCodeLogischer SchlussVertauschungsrelationTreiber <Programm>AnalogieschlussVersionsverwaltungProzess <Informatik>Automatische HandlungsplanungMAPFormale GrammatikAggregatzustandCodeProgrammbibliothekBitArithmetische FolgeProjektive EbeneZellularer AutomatAppletBildverarbeitungCodierungWort <Informatik>MultiplikationsoperatorInformationsspeicherungForcingDatenbankSoftwareURLMathematikSichtenkonzeptPatch <Software>MultigraphÄußere Algebra eines ModulsOpen SourceMixed RealityFormale SpracheMusterspracheBildgebendes VerfahrenElementargeometrieShape <Informatik>SummengleichungE-MailKorrelationsfunktionQuaderEreignishorizontOffene MengeQuellcodeComputeranimation
12:01
CodeMAPServerComputeranimation
12:19
EreignishorizontSoftwareentwicklerCompilerFormale SpracheGruppenkeimSoftwareentwicklerReelle ZahlBitFormale SpracheRechenschieberEreignishorizontCompilerVollständiger VerbandRechter WinkelMultiplikationsoperatorGruppenoperationInternetworkingMapping <Computergraphik>Sampler <Musikinstrument>Internet der DingeComputeranimation
13:18
EreignishorizontGruppenkeimWebDAVDivisionGruppenoperationBitURLMereologieVollständiger VerbandPunktwolkeCluster <Rechnernetz>DifferenteZentralisatorComputeranimation
14:10
SpywareServerElementargeometrieVersionsverwaltungMixed RealitySoftwareentwicklerEreignishorizontCodeSpywareHilfesystemProgrammbibliothekZentralisatorArithmetische FolgeComputeranimation
15:11
SpywareElementargeometrieServerVersionsverwaltungEreignishorizontZahlenbereichComputeranimation
15:38
E-MailMailing-ListeVersionsverwaltungMehrwertnetzBildschirmmaskeProzess <Informatik>Euler-WinkelEreignishorizontCAMSoftware Development KitURLE-MailTermProjektive EbeneProzess <Informatik>Raumfüllende KurveInternetworkingGruppenoperationEreignishorizontt-TestBitFunktionalMailing-ListeVersionsverwaltungMAPSystemzusammenbruchBitmap-GraphikMultiplikationsoperatorOpen SourceCodeStellenringGamecontrollerOrtsoperatorGüte der AnpassungGebäude <Mathematik>Web SiteAbstimmung <Frequenz>ProgrammbibliothekSoftwareHilfesystemSelbst organisierendes SystemSichtenkonzeptDämpfungTemplateBildgebendes VerfahrenAppletTextur-Mapping
19:16
MittelwertStellenringArithmetisches MittelEreignishorizontEreignishorizontWeb SiteVorlesung/KonferenzComputeranimation
19:47
Web SiteProjektive EbeneE-MailProzess <Informatik>Arithmetisches MittelAppletBildgebendes VerfahrenMailing-ListeVorzeichen <Mathematik>ComputerspielBeobachtungsstudieKontrollstrukturGruppenoperationBesprechung/Interview
21:03
GoogolRed HatSummierbarkeitElementargeometrieMailing-ListeBitmap-GraphikProzess <Informatik>Spezielle unitäre GruppeURLProjektive EbeneTelekommunikationHilfesystemZahlenbereichMereologieSelbst organisierendes SystemInstantiierungComputeranimation
22:08
SelbstrepräsentationHinterlegungsverfahren <Kryptologie>MereologieDatenstrukturURLFokalpunktVorlesung/KonferenzBesprechung/Interview
23:00
Computeranimation
Transkript: Englisch(automatisch erzeugt)
00:09
OK, apparently I'm wasting time. That's good because this is a developer kind of talk, so welcome. This talk is Getting It Done at Location Tech.
00:21
How many people have heard of Location Tech? Hooray! That's about half of you. For the other people, you'll learn a little bit about what Location Tech is, but mostly you'll learn a little bit more about our programming culture because, you know, programming is, you know, a fun activity.
00:41
So welcome to this talk. I'd like to introduce my colleague Tyler Battle. He's a software developer at Boundless North in Victoria. And Tyler has been joining me for a year, year and a half? Almost two. Almost two? Yep, it's November.
01:01
Wow. I'm shocked. OK. Tyler's been joining me for almost two years working on the GeoGig project, and it still hasn't made a release. No, it has not. Can you work harder? Yes. OK. We're almost there. Excellent. Excellent. Super close. Excellent. And my name's Jody Garnett.
01:20
After picking on Tyler, I'll let him introduce me. This is Jody Garnett, who is a community lead at Boundless. And he's working on lots of projects, including UDIG and JTS. And what other stuff are you doing? Oh, you're on the PSC. And I made you do this presentation.
01:40
Yeah, and you made me do this presentation. Yes, excellent. This is how you can tell I'm a community lead. I volunteer people. Excellent. So we work at Boundless. Boundless is one of the sponsors of this conference. And Boundless works with LocationTech. It's a member of the LocationTech working group.
02:01
And yeah, they pay us to do fun stuff like talk to you today. So while I'm really happy with Tyler, because who wouldn't be? We've got a couple of people who we couldn't be here today. And sadly, these were the people I was really interested in hearing from, because I hear from myself way too often, and Tyler as well, about once a week.
02:23
So Rob Emmanuel is from the GeoTrellis project. And James Hugh from CCRI is from the GeoMesa project. And these people are both kind of come from some of the cloud community projects that are taking shape at LocationTech.
02:44
And they've both been making a lot of progress this year. And I was really hoping to get a status update from them and hearing how it's going. One thing that's really fun about this is you see that they've got a couple of projects in common, like SFCurve. Now even though these people come from different organizations,
03:00
they're working at LocationTech, and they've been able to set up some joint projects together in order to share a library about space-filling curves. Because who would want to write space-filling curve twice? Yeah, so I'm really sorry they're not here. So LocationTech is a working group.
03:20
Developing advanced location-aware technologies. And it's part of the Eclipse Foundation. The Eclipse Foundation is most famously an ID for developers. But it is also a software foundation for people working on commercial-friendly open-source software, like us at Boundless.
03:42
Now, Eclipse is a not-for-profit, member-supported corporation. So here's what Eclipse looks like from our perspective as developers. It's kind of up towards the top of the slide and over top of this little hat thing. I think it's a mustache. It's a mustache? Okay.
04:01
And Eclipse has a number of working groups. So on things like science. Internet of Things. Polaris Sys. What's Polaris Sys? That's the embedded. Embedded? It doesn't say embedded. It does not say embedded. I have no idea why it's called Polaris Sys. And Eclipse Automotive, I can understand that. Yeah.
04:20
It's like you take your Eclipse and you drive it somewhere? Excellent. And LocationTech, which doesn't say... Well, it says location and technology, so that's good. But that's a pretty advertised... That's a good description. Here's our LocationTech members. IBM and Oracle and Boundless. I've heard of them somewhere before.
04:41
Red Hat and Google. And we've got a whole bunch of participating organizations. CardoDB, CCRI, where Jim is from. Vivid Solutions, Ordnance Survey. We didn't update the Cardo logo. It still says CardoDB. They changed their logo? Why would they do that? Because it's Cardo. Okay.
05:00
Sorry, Cardo. And then we've got a number of guest memberships. OSGO is a guest, and several universities. From our perspective, LocationTech is working for developers. This year, we actually have had our first project actually graduate and make a release.
05:26
So from a developer's perspective, as long as we're releasing, everything's okay. Made it through earlier this year, about January. It took a little while to figure out the technicalities of how to make a release. So since this was our first official release,
05:41
we had to figure out how to do things like how to sign the artifacts. So they said in the metadata that this was from LocationTech. And also, once we had everything built up, how to actually start the release review. So it could be reviewed and checked and be made an official release. But yeah, they sorted that out.
06:02
The GeoMesa project has also succeeded in making a release. And they made a lot of progress at one of these incubation sprints, which we set up. And they worked... Jim in particular has worked really hard to get this release out. This release is based on GeoTools 14.1. And one trick that Jim did was he combined forces with
06:21
UDIG and GeoGig and others in order to go through and review the GeoTools code base. GeoMesa project's been super active this year and is balancing official releases with these patch releases. Yeah, so the patch releases are a bit easier to put out. You don't need to go through as much formality for those.
06:42
So they've been doing a good job of keeping pretty regularly every few weeks getting patch releases out. So that's been really good. Okay. Another project that's been making a lot of progress is JTS or the JTS Topology Suite. Who here is like use JTS?
07:01
Who here is used like Geos? Who's used like QGIS or PostGIS or any open source spatial software ever? Excellent. It all comes down to this one project that implements geometry, which is kind of like the rocket science of our spatial open source industry.
07:23
So JTS has been ported near and far into different languages. And it's the geometry, it's the shape of our industry. One of the things is when we did an incubation sprint in Victoria,
07:40
all the other project leads, because we're so indebted to this project, kind of dropped what we were doing and we stepped in and helped this project. So we helped migrate it to GitHub. I think it was on SourceForge. Yeah, it was. SourceForge is still a thing? It is apparently still a thing. Okay. We also helped reviewing the headers because the projects re-licensing from LGPL,
08:01
which is terribly unfashionable, to BSD and EPL. We also helped migrate it from ANT to Maven because Maven is so 2003. And for our pains, we were rewarded with committer status. Now that said, the JTS project is making lots of progress,
08:22
but it's currently stuck on re-licensing. So we need to contact a few prior committers and say, yo, are you okay if we change from LGPL to these other things? So when you're starting a new project, it might be worth considering having agreements with the JTS project. Allow the project to re-license so you don't have to go through all of this.
08:45
Just thinking going forward. You know, if you want your project to last over 10 years. Yeah, yeah. Do you want to talk about GeoGig? Plan for success. Plan for success. So GeoGig is a project that does versioning for geospatial data.
09:03
It's been going for about two years now. And it's made a lot of progress recently, but we've been having some issues getting to a 1.0 release. We were using BerkeleyDB as our main data store, which is a problem because it has a somewhat dodgy license
09:22
that isn't quite open source. So we've been in a holding pattern waiting to get some alternatives using either SQLite or Postgres. And those should be ready quite soon. We have an RRC3, which just came out a week or two ago.
09:45
And that includes a bunch of new features, including some work on the Postgres background. And yeah, we should actually be getting pretty close. But one of the tough things is when you are going through this process,
10:01
you do have to go through all of your dependencies and make sure that there are no licensing issues. So sometimes it takes a while. And sometimes I would say there's a correlation between code quality and the level of licensing adherence. So sometimes it's good to have to go through and cull things every now and again.
10:23
In particular, the GeoGig process taking so long, the version of GeoTools changes every year, which means that Gabriel or Tyler needs to submit a new copy of GeoTools out. If you have any more interest in GeoGig, there is a talk tomorrow.
10:42
UDIG. So UDIG is a project which I'm the lead on. And we had a pretty good year. We had a great time at EclipseCon Europe. They had a location tech day. And we've been looking at a few things, such as making a cut down version with less dependencies. And we've also been combining forces with GeoMesa
11:01
to share some of this effort. One thing that's kind of needed here is we need a replacement for the Java advanced imaging image processing library we need because it's not open source. It kind of came out of the box with Java. So we're going to need to write ourselves a replacement. Or we might be able to treat it as a works with dependency.
11:21
So if you've installed this image processing library in your Java, UDIG might be able to pick it up. This works with approach is really helpful for things that are not open source, like database drivers, for example. Let's move on to finite location tech. We had a great time at the Philadelphia Code Sprint.
11:42
And this was an OSGEO code sprint with a really strong mix of OSGEO location tech projects and the cesium team. It's an OSGEO event sponsored by location tech and organized by one of our member companies at Azavia. And I went there to work on UDIG and spent most of my time doing GeoTools outreach.
12:02
So here's a bunch of cheerful people. There's Jim. There's Jim cheering. He looks very happy. Yeah. There's Rob in the corner working. There's the map server team in the other corner not working. Oh, wait. One of them's working. OK. We also had a really interesting, for us, conference,
12:24
the Phosphor G NA in 2005. So this was a joint EclipseCon Phosphor G event. And it was really fun to meet real Eclipse developers because we feel a bit like imposters because we're here to work on stuff. And I will fix the slide.
12:43
I'm fixing the slide right now. There we go. OK. One thing that was interesting is there were a lot of them. And not all of them were compiler language geeks. And we really had a great time hanging out with the science working group. They felt like our people. We also suspected that they had more fun at our talks
13:02
than we had at their talks. Yeah, a bunch of the internet things people were really excited about the mapping stuff too. Yeah. So it was a really great event. And here's an example of them having fun at a Planet Labs talk. We also have a location tech day.
13:22
And this was part of EclipseCon Europe. And I was really surprised and shocked at how much larger the Eclipse community is in Europe than in North America. While our talks went well, the highlight for me was kind of crashing the science working group meeting and seeing some of the fun stuff they're doing and catching up with Frank Gastorf,
13:41
who's one of the UDIG committers. The science working group was kind of fun. We've got all our cloud geospatial stuff going on. And Rob's talking about 1,000 node clusters. And the science working group's looking at him and says, one of my nodes has 1,000 cores in it.
14:00
So it was just a little bit of a different experience. So here's us kind of having fun, having beer. Beer is a form of fun. And another kind of central event for us was an incubation sprint. So we kind of hijacked people being in town for the GeoServer Wicket Sprint this last January.
14:23
And this was a longstanding request from our developer community. And it was only like three or four days, but everyone made so much progress, often by just having a chance to focus on what we were doing without all the distractions at work. JTS went from kind of like 10% or 20% done to 90% done,
14:44
just over the course of a couple of days and a lot of help. And GeoMesa really sat down and finalized most of their dependencies, really set them up for that first release. And GeoGig mostly went on a witch hunt. So it gave up things that weren't really very well done, like the osmosis library is kind of a mix between GPL
15:05
and LGBL code when we went in to look at the details. And UDIG, we did a lot of research. But yeah, that was a really helpful event because sometimes it's hard to get motivated to do some of the IP review. And it was really good to get everyone together
15:22
so that people who had done it before could mentor people. So JTS was amazing. It went from absolutely no work done to almost fully through in a couple of days. That's pretty good. Safety in numbers.
15:40
This is one of these rare pictures of you, Tyler, without a beer. I know. I know, yeah. This is where we took down GitHub. That was pretty good. So just for our final day, GitHub crashed. So we all had pink unicorns. It was the big JTS commit is what took down GitHub. I'm sure of it. I'm sure of it.
16:02
OK. Just a bit of how GitHub functions. There is a technology PMC top level project. What does that actually mean? Well, the top level projects kind of are used to group things. So as location tech grows, we might split this up into like libraries and processing and things.
16:20
What does this actually mean? It means that the leads of the software projects have a hangout once a month, and we see how we can help each other out, and we send the notes to the email list. That's been really strong community building activity. In terms of helping people out, it might be something like GeoMesa wanting to use a new version of a cumulo,
16:42
and I could check that out and confirm it was open source and stick a plus one on it. The other kind of things we do is we do new committers and so on, and this is actually done not with like an email list. It's actually something that's kind of automated on the Eclipse website, and it's a little bit tricky.
17:02
You can kind of go through this website and have the voting approved and have the steering committee approved, but we also need to ask the candidates to sign some paperwork, that contributor license agreement kind of thing you mentioned earlier. So this is where JTS ran into trouble because the lead Martin didn't have control
17:22
over the code base. He wasn't in position to change the license very much, and that's kind of good. The other thing we do is we end up starting new projects. So we saw that space-filling curve project start, but you can go through here if you're looking at joining location tech
17:40
or if you're looking at starting a new project from scratch, you fill in a form, you send an email to emo at eclipse.org, which I have no idea what that stands for, Eclipse Management Organization. You'll have lots of acronyms to learn. Yeah, for a creation review, and that kick-starts a process
18:01
where they check to see if there's any trademarks. You might have to answer a few questions, and at the end of that, they provision you with infrastructure, but we mostly all use GitHub. Yeah, so I had the pleasure of submitting the raster processing engine, which I hope will be a replacement for Java Advanced Imaging.
18:22
I submitted that last week, and I was hoping it would get approved this week. Public now. Excellent. That's very exciting. Late-breaking. You can read that thing on the internet. In terms of outreach, location tech does all kinds of outreach from sponsoring events like this.
18:41
One activity we really enjoy is the location tech tour, which rather than have a big conference takes the speakers on almost like a road show, visiting smaller groups. We really enjoyed a location tech tour stop we did at the local Victoria University. It was really fun to be in an ESRI Centre of Excellence
19:04
introducing students to open source for the first time. Any questions? If you don't have questions, I'll just stare awkwardly at Tyler, and we'll see what happens. Hey, everyone. Hey, Jody. This is a comment, not a question.
19:22
You may splash up the tour site with eight events. Thea, let me know that there's going to be at least 15 events this year. And if anyone here is interested in having an event in their town, reach out to myself, Andrea Ross, or Thea Aldrich. We're here to help, and we'd love bringing the tour events to new cities. I would have put that contact details,
19:41
but this was a dev talk. Devs can organize things. Devs can organize things sometimes. First of all, I'm very happy that you're working or that you want to work on the Java Advanced Imaging problem. I think I just surged back for it.
20:02
I started complaining about it six years ago on a mailing list. Anyway. Wait, does that mean you're volunteering? No, I don't know. No, but I wanted to say the opposite. You saw my job title, Community Lead. I volunteer you. What I wanted to ask is,
20:21
is there anything public where I can point people to? Because I know there are a lot of other non-GIS projects who actually have the same problem. So we are really setting it up to not be specifically geospatial. And we're doing that by interacting with the Science Working Group and other interesting parties.
20:40
And Andrea Ross just mentioned that the little project proposal is now public, and you can see it on the Eclipse website. Okay, so I can point them there. If I was especially brave and handed this mic to Tyler here, I could go to the website. Uh-oh, we're going live. I'm worried.
21:10
While we wait, are there other questions?
21:25
Just out of curiosity, since I know that you work with both organizations, Jody, I know that it seems that because there are fewer projects currently on location tech,
21:42
that there seems to be very tight-knit, which is lovely. Is that, I mean, correct me if I'm wrong, but it sounds like there's a little more communication between the projects because it's just a fewer, I mean, you know. There is a smaller number of projects, and often we are kind of learning the ropes together. It's also very well set up
22:01
to help mentor project teams that are new to open source. And so I quite value that experience. It's one of the reasons why I take part. I should also point that the structure of location tech is organized with members on a steering committee, but the committers do have a voice.
22:22
And so I am one of two steering committee representatives, sorry, two committer steering committee representatives on that steering committee. So the committers, even though they're not paying members, they do get a say in what's going down.
22:41
Any questions? Thank you guys very much. It's so nice to do like a dev-focused talk on location tech. So I thank you all for the opportunity to speak.