Merken

Push it through the wire! Push it more, if it's wireless!

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
the
better OK the phone hiding my name is so you hinted it sky welcome to my talk about how to get so they just wanna or
humorless I have been involved uh longer than 3 years 10 10 years have the development of open source of fur for geospatial especially I studies is on the scope side using gradually developing 4 and with crowds chairs and isolate moved to server side and last year's likable fewer Simon I make my living because they do make my living through a development of just replace that G. I. is obligations whatever we what is
this I have OK you probably know that
the development of but based applications has completely changed through during the last couple of years and they are no longer writing just get jokes and sending it somehow directly to the client signed like it it's in the early days this is nice
yeah you the I from here and I each set up to the presentation I
checked the 1st slides there seem to be retained and now as you may see stress like nothing the concerning you all the way compares to 104 I see thing to OK thank you for the link analysis OK sorry and others that I'll have to the move faster thank you and and and trained and and the and the and it was yeah the FIL
try and I
yeah full works been yeah so
we are no longer sending Joe script file so that the browser and
instead we are compiling the coach again the good and faster of the transfer that so many transformed into the quantitative program itself isn't that big and you can compress for example code of OpenLayers 3 right now from more than a megabyte then several kilobytes for comp concrete example on the other side this is what we are facing a at least team I'm working with the compilation of the whole obligations extracts is and if you are deploying and and as you won't have continuous integration it takes you really a couple of minutes so the say maybe on our or 2 a day that together called compiled yep the but now these are just LR understanding the printers added they don't have suburban which rendering really big amounts of data which could we could see in the fall of the presentations of during the whole force for not only forging and roster images them is pretty simply are coming with compressed in a compressed form using Jeep extra PNG files and for rendering thruster image there's authorization need is anymore actually because at the end of the finally somebody and you have to transfer data from the vector thrust of course in order to be able to display them the and the amount of information in is a file can be significantly lower compressed of 2 extra files now we consider the amount of information of interest to file this of course limit is to its resolution and number of course you are able to somehow to transfer to the user so the user can really still interpreted what amount of information in a vector file there is given by so the a number of vertexes you you units for description of a future like for example the dog and and uh of course number of attributes you are able to put into the to intuitively to
have from certain point of so far our differences in level of detail that size of a victory might be bigger actually then the size of prosody fellow
and why would you need to that at the scope of this presentation is on disaster for vector related to the to the clients and why would you need is an anyway right and how are you want to do users so to enable at the inter entity interaction metadata they're like dynamic maps you you want to some whole distributed dynamic aspect of it are you of course you want to enable the user to edit the data and more other reasons but the main reason why I think it's still the most important 1 and that is that vector teeth are simply cool and this gives you all the power of having the data on client side you can you really whatever you need to do with them but we are faced facing here the common database for the money you can't have all 3 of the views of the data at once uh even you if you keep trying to keep pushing whatever you try to do you always so have to struggle either you will have beaks of the data or you will never delivered at a time of fast way order data will be up to date you source is always just 2 of those cells use but never all 3 of them whenever you try to do do those of suggested steps when you deal with a large amount of data and and of course a victory the revise strategy so say and it starts from the beginning from the very the simplest 1 and because the most complicated world uh the 1st Scott step ace compass the communication between the server and the client then the other choose a proper part file for Chile for the data transfer itself is cash obviously use lossy compression for victory to as well you can do business Russell's 104 actors and transfer only the data of user literally means to see your application really needs to have to behave because of some in some weighted uses the feeling of having to rethink and the 1st step compresses service server-client communication and might be obvious right but you deeper was surprised the whole from the system administrators for gets the most simple the simplest think switch on for example multi-flight on Apache Server than that if you if you take megabytes big judges and file its companies like 50 thousand police polygons have based on opposite can be done you can really make a 10 times smaller just and just like that just with single configuration option so don't know don't forget this this is probably the 1st slow start if if something is slowly feeling there's something slow this would be probably the 1st check out the other the other and check would be you I use the proper file format everybody's talking about you Jason right well based but surprisingly to me as is really big surprise Judaism of isn't the most compressive compressor a fair amount after victory tell if you compare to GML surprisingly if slice and bake again and speaking about 50 thousand of polygons and based on OpenStreetMap data randomly picked up from the deficit I am not sure why this is possible I have rechecked several times so I have a feeling that the DG among file doesn't contain of tributes about where there's a will you wild Joe Jason at through literally all abuse really believe in the little elegant I think so yeah and the I just use some that's OGR library 4 MIT transformation the OK so but on the other hand of course as as isn't everything I you can imagine parsing 48 megabytes GMO and the browser takes a certain amount of energy and processor time compared to to judges so as well what I want to say just be use cash and
other strategy the cache doesn't necessarily have to be you don't even have to set up a little dialog Cash Stack button might be enough if you turn on for example that the framework can if you're if you're using and that the framework for recording or whatever and you simply can't just a little amount of your data on the public and the server side so that the application doesn't have to really touch the database something has changed is the reason for touching the database and if you want to deliver data really in a fast way and for some reason might be data size might be were not badly optimization and Butler index files In polls jazz atabase whatever if you turn on cash and application server it makes it for a really really much faster then and if you go to that data directly and again it doesn't cost you basically everything's just few lines of code the and so you can have a fast result actually were easily
OK and and such AG and if you are displaying compressed it Beck J. pick files for example and f for users and you you everybody knows that's how it looks nearly as the relative but users of complaints is good enough for them so why couldn't we do the same with vectors and already mentioned here as for as well it's
actually 1 thousand the vertexes Berlyne for of course to attain zoom level and when for example 3 vertexes or for would be enough for the data line or polygon or whatever representation the from pink libraries and it more than once like open source open 3 of causal orally flat they do we actually have a preprocessing of the data which are at hand before they surrender rates in for example conoscenza if so there really then that and most expensive part rendering itself this is using this working This really a compressed along the number of vertexes if you are a new wrinkle lines were braille's but of course but that doesn't solve the problem how to and it that you need to know the data to be transferred to climb client this is everything happening just on the client side with and all this but I I would advise you to have a look at the application would there be can plates sets a number of 5 exists unique prime and to see what's happening there and this is this is happening condo on the on the client side somehow in the because the 1 of the former that usually implements was accomplished compression is to produce and former that it a origin on Joe Jason and again my concrete example if I take any 2 megabytes 2 adjacent polygons I've converted to to produce song it has only 3 my God and how is it it's using monte flight this is a logistical kilobytes it they seem to be magic to me and I would like someone in the audience to reproduce such as experiments so this because it seems impossible but it rendered it has written a book so like supposed to be and they just a little bit at the proposer you you can get pretty good results the but is this introduced by my boss torque everybody knows the produce then like think it doesn't make sense for much further amended for 1 reason that I don't really understand what's behind its what's behind the magic except Fourier it uses throughout coordinates values use that are from some starting points with the city and there is a internal index of arcs of so called and so there is a the link yeah there's a link to documentation on get out where you can get into detail I wanted to make a picture had OK the prison the should be only OK and uh as I said if you this list a compression since and the delta values are integer values basically so you you really don't you you lose a set and a level of detail that for most of the applications is far enough but there is who was raised as facing with the producers the support for so called bit a bit integer numbers which again should be a program for 99 per cent of your use cases that I my use case is I want to display possible pedestal possible data into a browser and today D is bake urine then of course there's a workaround you know you can create your own ities and you strength for example for this 1 book that it isn't it plans and I wanna go our next steps would be you can you don't have to transfer to you don't want to transform the data which are hand in 1 step from the server to client so tonight we're speaking commendatore shrinking but now the and the strategies for the user is feeling to that application is somehow living that are laid out there you know they might be coming in the house all steps the you don't have to those are I think in 1 step yeah useful piled rest 1st and everybody speaking to bold piled vectors the the and over several talks already have altered our public sources and in this country's example is based on openness and it is actually adjacent adjacent and tireless which a change and others 5 minutes our the it's relatively complicated to get them and compatible to clusters of course but they're not this libraries they supported to with basic of problems what you can see here in the presentation of chicken the great and is so this editing tool Strauss about that this was a bucket basically an open Madison discover so it's a problem that then you need to set up the whole
infrastructure that make things with a bit complicated whenever possible solution would be the final stage the project and biking this project which generates you which is able not only the genetics of us but also GOG sometimes and even the produce tiles which is very nice and I was
facing facing some some difficulties using
different projection density comic outsourcing will move them counselor or that McGovern projection but that was the only Jew I hadn't unrusted works nicely it
has a standard is met book styles again we're addressing this or during the Office for she met my books as the developed and opened a standard for basically a binary binary from vector encoding interesting case that the data are in accordance of the data do not use the geographic according its anymore it uses the corners of the image or the whole month so rendering calculations this like nickel collections of data and is pretty fast and of course IT supported leaflets How can optimize it's going to be supported so means that I really tested it on my computer it's in a branch and it will be there so and for generation you don't have to needs and to use my books infrastructure you can of course you can use larger server or again in the very near future map server I sort how there might be another strategy you might still want to take if you if you have a look at our disposal and apparently displayed and this case and 2 4 4 different vector features and it is further like it this in them in in europe map application you will have at least 4 different features with the same ID for example and then if you provide some filtering some selections of course obligations to be around adjusted to to the fact and another option would be to to download just a vector data which for example centroids fits into a particle tile right now the
that it could look like this you see this down clear boundaries but because under endemic whole lecture features which centroid sorry which censored fit it
into the virtual so to say which 1 in books of the time and of course if you have some long a power source polygon-data data for example also want formed at present . it up on the edge of dozens of dozens title then add feature won't be loaded the best but for most use cases this seems to be a competitor well and then you have only 1 teacher for 1 feature 1 times as 1 object in your application
yes and result of this effort and only at and a user will see just what he or she is expecting and when you adding you've thruster chrysotile so that means there will be somehow loaning tiles them that those would appear and use of a feeling that something is happening which is always good the conclusion the the 1st question that I would be you want to know how use all this compression 5 4 months totaI link static when I want to display a single GPS look for an answer so this all and those approaches are just for situation in application seems to be slow for some reason if you're trying to look some because amount of data and user-definable links and of and yeah so those would be a substitution you can take if you want to reduce the world but so the time to time and time and rendering time the caching and styling of it gives you more speed but of course you always can have only to attitudes of the thing the and ricocheting you rules so timeliness of the data because there will always be some cash the from some kind of application is perfectly OK everybody can live is Google they time-window everybody knows that p cashed and you never get the same result asking from different uh outside of the rope asking will on or if you have you never get the same result and tool parts of world in the same time because they databases isn't centralized and so on down of up-to-date no that's fine good enough the different Indonesian regressions few so essentially if like in 5 minutes and 2 min. regressions so 1 for true yeah Vincent please OK like is like so you did try to provoke as to the matter of estimated and yes it works have alleviated just their today together of its arms must move forward and under the sun the while from exercising the other 1 from your below site and that my data by data means really really cadastral data could possibly to and to that of shows that the database and I wanted to show it but I wasn't able to come about that my computer into the into this PGA output uncertain so I can show it to you later surprise in private the OK thank you then I would like to ask next speaker a friend
Beobachtungsstudie
Open Source
Server
Softwareentwickler
Drahtloses lokales Netz
Computeranimation
Client
Kartesische Koordinaten
Computeranimation
Rechenschieber
Open Source
Freeware
Binder <Informatik>
Kombinatorische Gruppentheorie
Normalspannung
Computeranimation
Analysis
Skript <Programm>
Elektronische Publikation
Computeranimation
Bitmap-Graphik
Router
Subtraktion
Punkt
Compiler
Zahlenbereich
Wärmeübergang
Code
Computeranimation
Übergang
Deskriptive Statistik
Bildschirmmaske
Knotenmenge
Einheit <Mathematik>
Typentheorie
Code
Inverser Limes
Optimierung
Bildgebendes Verfahren
Bildauflösung
Attributierte Grammatik
Autorisierung
Vektorgraphik
Kontinuierliche Integration
Sampler <Musikinstrument>
Übertrag
Vektorraum
Elektronische Publikation
Rendering
Forcing
Verschlingung
Ablöseblase
Information
Ordnung <Mathematik>
Resultante
Browser
Minimierung
Program Slicing
Kartesische Koordinaten
Computeranimation
Metadaten
Client
Eigentliche Abbildung
Mixed Reality
Quellencodierung
Gerade
Caching
Sichtenkonzept
Prozess <Informatik>
Datenhaltung
Quellencodierung
Quellcode
Dateiformat
Konfiguration <Informatik>
Verband <Mathematik>
Framework <Informatik>
Rechter Winkel
Automatische Indexierung
Server
Strategisches Spiel
Client
Dateiformat
Ordnung <Mathematik>
Telekommunikation
Server
Interaktives Fernsehen
Zellularer Automat
Keller <Informatik>
Vektorraum
Transformation <Mathematik>
Kombinatorische Gruppentheorie
Polygon
Code
Framework <Informatik>
Programmbibliothek
Coprozessor
Konfigurationsraum
Leistung <Physik>
Vektorgraphik
Systemverwaltung
Browser
Einfache Genauigkeit
Datentransfer
Telekommunikation
Gasströmung
Vektorraum
Elektronische Publikation
Mapping <Computergraphik>
Energiedichte
Polygon
Caching
Mereologie
Eigentliche Abbildung
Resultante
Bit
Punkt
Mengentheoretische Topologie
Browser
Selbstrepräsentation
Kartesische Koordinaten
Ähnlichkeitsgeometrie
Computeranimation
Übergang
Kreisbogen
Client
Wärmeübergang
Quellencodierung
Gerade
Präprozessor
Quellencodierung
Zoom
Bitrate
Gefangenendilemma
Menge
Automatische Indexierung
Ganze Zahl
Benutzerschnittstellenverwaltungssystem
Grundsätze ordnungsmäßiger Datenverarbeitung
Strategisches Spiel
Server
Mathematisierung
Zahlenbereich
Sprachsynthese
Maßerweiterung
Kombinatorische Gruppentheorie
Polygon
Knotenmenge
Ganze Zahl
Programmbibliothek
Cluster <Rechnernetz>
Optimierung
Open Source
sinc-Funktion
Mailing-Liste
Vektorraum
Elektronische Publikation
Binder <Informatik>
Kreisbogen
Zeichenkette
Moment <Stochastik>
Polygon
Offene Menge
Mereologie
Räumliche Anordnung
Server
Bit
Schreib-Lese-Kopf
Parkettierung
Vektorraum
Projektive Ebene
Natürliche Sprache
Computeranimation
Subtraktion
Server
Vektorraum
Kartesische Koordinaten
Computer
Fastring
Binärcode
Computeranimation
Trennschärfe <Statistik>
Bildgebendes Verfahren
Verzweigendes Programm
Vektorraum
Rechnen
Office-Paket
Dichte <Physik>
Konfiguration <Informatik>
Mapping <Computergraphik>
Generator <Informatik>
Parkettierung
Server
Strategisches Spiel
Lesen <Datenverarbeitung>
Projektive Ebene
Partikelsystem
Standardabweichung
Unendlichkeit
Objekt <Kategorie>
Randwert
Kartesische Koordinaten
Computeranimation
Fitnessfunktion
Resultante
Datenhaltung
Verhandlungs-Informationssystem
Quellencodierung
Vektorraum
Einfache Genauigkeit
Schlussregel
Kartesische Koordinaten
Computer
Euler-Winkel
Binder <Informatik>
Dateiformat
Computeranimation
Datensichtgerät
Open Source
Freeware
Parkettierung
Caching
Lineare Regression
Mereologie
Vorlesung/Konferenz
Substitution
Quellencodierung
Funktion <Mathematik>

Metadaten

Formale Metadaten

Titel Push it through the wire! Push it more, if it's wireless!
Serientitel FOSS4G Seoul 2015
Autor Cepicky, Jachym
Lizenz CC-Namensnennung - keine kommerzielle Nutzung - Weitergabe unter gleichen Bedingungen 3.0 Deutschland:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen und nicht-kommerziellen 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 und das Werk bzw. diesen Inhalt auch in veränderter Form nur unter den Bedingungen dieser Lizenz weitergeben.
DOI 10.5446/32098
Herausgeber FOSS4G
Erscheinungsjahr 2015
Sprache Englisch
Produzent FOSS4G KOREA
Produktionsjahr 2015
Produktionsort Seoul, South Korea

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Today's web browsers, their rendering engines and JavaScript interpreters are able to display relatively big amounts of vector data. Moving from DOM rendering (as it was implemented with help of SVG in for examples OpenLayers 2) to Canvas (and further to WebGL -- as we are now having in OpenLayers 3 or Leaflet) enables us to display thousands of complex vector features, with complicated on-client vector data styling. With this possibility, we are facing now new types problems: how to send such amount of data through limited internet connection? If we have closer look at the problem, we can see clearly, that old database paradigm has raised one more time: we can not have all three attributes of data in one pot, but only 2 of them: speed of the delivered data or amount of delivered data or their topicality. If we take this limits into account and decide to deal with big amounts of data in fast way, topicality must be sacrificed. In the talk, we will demonstrate some possible solutions for this problem, using tiled vectors, generalization, aggregation of vector data. Also advantages, disadvantages of various new and popular vector formats, such as GeoJSON, TopoJSON or MapBox will be discussed. Geometric data do not have be rendered all the time in all scales and over whole area of interest, but only necessary portion of them. If displayed in smaller scales, aggregation and generalisation can take place on the server side. That implies, that using vector caching mechanism could be considered as well. But if we need direct interaction of the server input with cached vector data, mechanism for this must be defined as well. Also attribute data have to be transfered separately, if all the optimisation was put in the vector geometries. Also possible steps between cached data and real-time data will be discussed.

Ähnliche Filme

Loading...