Slimming down OpenLayers

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Slimming down OpenLayers

Arnulf Christl
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Folks,
pertinent to the discussion yesterday about where the project is going
Christoph again proposed to go for OpenLayers as the (only) viewing
component to reduce the scope of development. One argument that I always
bring up against this is that OpenLayers is way too fat. It invariably
crams several 100k through the pipe and into the browser while all that
is needed in many cases are just a few 10k. With 2.9 this has not become
any better. Today I found this blog [1] that describes hwo to reduce the
size of OpenLayers from a user perspective. If we go for OpenLayers
could we work backwards by starting with as little OL code as possible
and increasing as we add functionality? Then we could always trade off
functionality for weight and balance both as best required by the
corresponding project.

Best regards,
Arnulf.

[1] http://linfiniti.com/2010/05/slimming-down-openlayers/

- --
Arnulf Christl

Exploring Space, Time and Mind
http://arnulf.us
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkvyUnYACgkQXmFKW+BJ1b3CowCfafEDWDNZatU7CJM0583iJSWz
cqoAn1rlluTyK5NH7WxDDx/zhlGuUA/C
=HJoM
-----END PGP SIGNATURE-----
_______________________________________________
Mapbender_dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapbender_dev
Reply | Threaded
Open this post in threaded view
|

[Mapbender_dev] Re: [Mapbender-dev] Slimming down OpenLayers

Christoph Baudson-3
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thanks for bringing this us.

you can easily load only the components you need. There are tools like

http://openlayerer.appspot.com/

on the one hand, but Mapbender would do the build itself. You would have
an Openlayers core module, and load everything else you need in
application elements. It would be like a custom build.

I think you can get Openlayers at roughly 100K this way.

Christoph

Seven (aka Arnulf) wrote:

> Folks,
> pertinent to the discussion yesterday about where the project is going
> Christoph again proposed to go for OpenLayers as the (only) viewing
> component to reduce the scope of development. One argument that I always
> bring up against this is that OpenLayers is way too fat. It invariably
> crams several 100k through the pipe and into the browser while all that
> is needed in many cases are just a few 10k. With 2.9 this has not become
> any better. Today I found this blog [1] that describes hwo to reduce the
> size of OpenLayers from a user perspective. If we go for OpenLayers
> could we work backwards by starting with as little OL code as possible
> and increasing as we add functionality? Then we could always trade off
> functionality for weight and balance both as best required by the
> corresponding project.
>
> Best regards,
> Arnulf.
>
> [1] http://linfiniti.com/2010/05/slimming-down-openlayers/
>
_______________________________________________
Mapbender_dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapbender_dev

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkvyWV8ACgkQGtMIfbycMX7RUACbBBEIYHf7ks38nnc2+z2myt/X
88QAoIxCO6HrBEx4IttTLkxEIwmTA7GL
=Jb8e
-----END PGP SIGNATURE-----
_______________________________________________
Mapbender_dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapbender_dev
Reply | Threaded
Open this post in threaded view
|

Re: [Mapbender_dev] Re: [Mapbender-dev] Slimming down OpenLayers

Thomas Baschetti-2
So what does this mean?
Do we loose functionality when focusing on openlayers (only)? Shall we do
this for 3.0 or in a subsequent release? What needs to be done?
Personally i really like the idea of just taking openlayers as the
mapcomponent,
and i don't care about some more k of code which have to be pushed
through the lines,
but of course there may be other opinions....


Thomas


On 18.05.2010 11:09, Christoph Baudson wrote:

> Thanks for bringing this us.
>
> you can easily load only the components you need. There are tools like
>
> http://openlayerer.appspot.com/
>
> on the one hand, but Mapbender would do the build itself. You would have
> an Openlayers core module, and load everything else you need in
> application elements. It would be like a custom build.
>
> I think you can get Openlayers at roughly 100K this way.
>
> Christoph
>
> Seven (aka Arnulf) wrote:
> > Folks,
> > pertinent to the discussion yesterday about where the project is going
> > Christoph again proposed to go for OpenLayers as the (only) viewing
> > component to reduce the scope of development. One argument that I always
> > bring up against this is that OpenLayers is way too fat. It invariably
> > crams several 100k through the pipe and into the browser while all that
> > is needed in many cases are just a few 10k. With 2.9 this has not become
> > any better. Today I found this blog [1] that describes hwo to reduce the
> > size of OpenLayers from a user perspective. If we go for OpenLayers
> > could we work backwards by starting with as little OL code as possible
> > and increasing as we add functionality? Then we could always trade off
> > functionality for weight and balance both as best required by the
> > corresponding project.
>
> > Best regards,
> > Arnulf.
>
> > [1] http://linfiniti.com/2010/05/slimming-down-openlayers/
>


--

Thomas Baschetti - Systemanalyse Geographische Informationssysteme
Hakenstraße 8D
49074 Osnabrück

Tel: 0541 25 91 90 | mobil 01577 189 25 91
E-Mail: [hidden email]
www.thomas-baschetti.de
Ust-IdNr.: DE264355072


_______________________________________________
Mapbender_dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapbender_dev
Reply | Threaded
Open this post in threaded view
|

Re: [Mapbender_dev] Re: [Mapbender-dev] Slimming down OpenLayers

Michael Schulz
Hi,

using OL as the main map component is a good idea. It will provide
Mapbender with a state-of-the-art viewing component. But I think we
can only take advantage of a part of OL functionality, since we are
only dealing with OGC stanardized data formats. Also, tiling is in the
WMS world not too widespread, so single tile mode still yields the
best results for not tile-prepared wms.

I think it makes perfectly sense from a cleaner-code and reuse
existing libs perspective, from a functional perspective I'm not sure
whether we gain a lot. It will take some time to have the full
functionality of the tree, digitizing and map interaction we now have
back.

We may also profite from the broad OL developer community, maybe even
toss some over to mapbender, because we are then providing a db
backend...

So, I think it makes sense to take this road.

Cheers, Michael


2010/5/18 Thomas Baschetti <[hidden email]>:

> So what does this mean?
> Do we loose functionality when focusing on openlayers (only)? Shall we do
> this for 3.0 or in a subsequent release? What needs to be done?
> Personally i really like the idea of just taking openlayers as the
> mapcomponent,
> and i don't care about some more k of code which have to be pushed
> through the lines,
> but of course there may be other opinions....
>
>
> Thomas
>
>
> On 18.05.2010 11:09, Christoph Baudson wrote:
>> Thanks for bringing this us.
>>
>> you can easily load only the components you need. There are tools like
>>
>> http://openlayerer.appspot.com/
>>
>> on the one hand, but Mapbender would do the build itself. You would have
>> an Openlayers core module, and load everything else you need in
>> application elements. It would be like a custom build.
>>
>> I think you can get Openlayers at roughly 100K this way.
>>
>> Christoph
>>
>> Seven (aka Arnulf) wrote:
>> > Folks,
>> > pertinent to the discussion yesterday about where the project is going
>> > Christoph again proposed to go for OpenLayers as the (only) viewing
>> > component to reduce the scope of development. One argument that I always
>> > bring up against this is that OpenLayers is way too fat. It invariably
>> > crams several 100k through the pipe and into the browser while all that
>> > is needed in many cases are just a few 10k. With 2.9 this has not become
>> > any better. Today I found this blog [1] that describes hwo to reduce the
>> > size of OpenLayers from a user perspective. If we go for OpenLayers
>> > could we work backwards by starting with as little OL code as possible
>> > and increasing as we add functionality? Then we could always trade off
>> > functionality for weight and balance both as best required by the
>> > corresponding project.
>>
>> > Best regards,
>> > Arnulf.
>>
>> > [1] http://linfiniti.com/2010/05/slimming-down-openlayers/
>>
>
>
> --
>
> Thomas Baschetti - Systemanalyse Geographische Informationssysteme
> Hakenstraße 8D
> 49074 Osnabrück
>
> Tel: 0541 25 91 90 | mobil 01577 189 25 91
> E-Mail: [hidden email]
> www.thomas-baschetti.de
> Ust-IdNr.: DE264355072
>
>
> _______________________________________________
> Mapbender_dev mailing list
> [hidden email]
> http://lists.osgeo.org/mailman/listinfo/mapbender_dev
>



--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Aufwind durch Wissen!

Qualifizierte Open Source Schulungen bei der
http://www.foss-academy.eu/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-----------------------------------------------------------
Michael Schulz
[hidden email]

in medias res
Gesellschaft für Informationstechnologie mbH

Schwimmbadstraße 2
D-79100  Freiburg i. Br.

Tel:  +49 (0)761 705798-102
Tel:  +49 (0)761 705798-0
Fax: +49 (0)761 705798-09

http://www.webgis.de / http://www.zopecms.de
--------------------------------------------------------------
Geschäftsführer: Stefan Giese, Dr. Christof Lindenbeck
Eingetragen im Handelsregister HRB 5930 beim Amtsgericht Freiburg
_______________________________________________
Mapbender_dev mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/mapbender_dev