[ka-Map-users] ka-Map and OpenLayers
Daniel Morissette
dmorissette at mapgears.com
Wed Jun 21 08:04:15 EDT 2006
Steve Lime wrote:
>
> One thought might be to look at tile-caching as a part of MapServer proper.
> Other projects (GeoServer, MapGuide, etc...) could follow suit... That could
> provide a clean break between client and server.
>
I think I'd rather see tile caching as a layer on top of MapServer and
possibly other servers instead of embedding it in MapServer, and let
MapServer concentrate on what it's best at: render high quality maps.
A simple way to achieve that would be to modify ka-Map's tile.php to
issue WMS GetMap requests instead of calling MapScript's $map->draw()
when rendering metatiles. This would allow plug and play integration
with any WMS server.
Down the road I think tile caching needs to become more advanced than
just the current tile.php and might need to become a beast in itself
that comes with its own http server (or as a fcgi or Apache module), at
that point, tight integration inside MapServer would probably make even
less sense.
Just my 0.02$
Daniel
> Steve
>
>
>>>>Paul Spencer <pspencer at dmsolutions.ca> 6/20/2006 4:18:23 PM >>>
>
> Good point ... we were discussing tiling today. I can actually see
> ka-Map turning into a tile-cache server project separate from
> OpenLayers. The shortest path to integration is to add a kaMap layer
> and tile class to OpenLayers and use the existing kaMap tile.php to
> serve tiles to it. I'm currently working on designing a better
> strategy for tile caching since the current directory structure is
> very limited for large caches (it only goes one directory deep when
> you really look at it).
>
> One of my personal goals would be to see a standards-based tile
> service come into existence. There has been some discussion on this
> on the webmap-discuss at mail.osgeo.org list and in the following wiki
> page:
>
>
>>http://wiki.osgeo.org/index.php/WMS_Tile_Caching
>
>
> I'm not sure I like the direction it is currently going, but I like
> the overall concept.
>
> Cheers
>
> Paul
>
>
> On 20-Jun-06, at 5:00 PM, Steve Lime wrote:
>
>
>>What would be the plans on the server-side? OpenLayers doesn't seem to
>>cover that aspect.
>>
>>That said, personally I trust your judgement w/regards to ka-map
>>and will
>>end up using whatever comes out of this union. The sooner a stable API
>>is in place the better since fear of shifting sands may keep folks
>>away or
>>from going into production.
>>
>>Steve
>>
>>
>>>>>Paul Spencer <pspencer at dmsolutions.ca> 6/20/2006 2:56:05 PM >>>
>>
>> From my initial review of the code base and project objectives, I
>>feel that it offers a cleaner code base and architecture. It is more-
>>or-less what I would do with the ka-Map code base if I wrote it.
>>
>>The differences are more in the project objectives. One of these is
>>to be able to integrate Google, VirtualEarth, Yahoo and other
>>commercial/free tiles without violating their terms of use. The end
>>result is that the application adapts to the limitations of an
>>existing set of tiles (tile sizes, projection etc). It is then
>>possible to build overlays that morph based on the underlying layer
>>source. I'm not sure this is a big priority for me, but it does make
>>the API itself more compelling I think.
>>
>>I'm not sure that I have described this very well. Perhaps John
>>could explain this concept better (since he was the one who explained
>>it to me :))
>>
>>John and I agreed that the projects should merge. He/OpenLayers are
>>willing to start from ka-Map and integrate the OpenLayers stuff if
>>necessary. I just feel that it would be easier to move the ka-Map
>>code into OpenLayers rather than the other way around.
>>
>>One big disadavantage to the ka-Map community is that the API will
>>change. We may be able to provide a compatibility API but I think
>>that would be more work than it would to help people convert their
>>apps.
>>
>>Perhaps a reasonable compromise would be to develop a Google API
>>clone (something I have been against previously), which could make
>>OpenLayers a compelling environment for existing Google-hackers plus
>>give them the option to move from Google in the future if the terms
>>of use/ads become cumbersome?
>>
>>Cheers
>>
>>Paul
>>
>>
>>On 20-Jun-06, at 3:15 PM, Steve Lime wrote:
>>
>>
>>>What does OpenLayers offer that ka-map doesn't ( besides a better
>>>name ;-) )? The
>>>list you reel off of limitations is pretty substantial.
>>>
>>>Steve
>>>
>>>
>>>>>>Paul Spencer <pspencer at dmsolutions.ca> 6/20/2006 1:28:19 PM >>>
>>>
>>>Hello ka-Mappers!
>>>
>>>I had a very interesting discussion with the developers of
>>>OpenLayers, a javascript api for tiled maps. Turns out OpenLayers
>>>and ka-Map are very similar. We decided that it would be beneficial
>>>for both communities if we attempted to merge the projects. The
>>>purpose of this email is to present this idea to the ka-Map community
>>>in order to generate some discussion on the subject.
>>>
>>>OpenLayers is hosted at http://www.openlayers.org/. They have some
>>>very interesting objectives which I think are compatible with ka-
>>>Map. Take a look at their site.
>>>
>>>OpenLayers is written on top of prototype.js, which was one of the
>>>things I was planning to do with ka-Map in the next couple of weeks.
>>>It has a very nice looking architecture and the code base is very
>>>readable.
>>>
>>>It is lacking quite a few things that ka-Map has:
>>>
>>>* core tiling engine is very similar, but it is not quite as optimal
>>>(doesn't reuse images for instance)
>>>
>>>* overlay stuff is point/text only
>>>
>>>* lacking tools (layer controls, scale bar etc) and the windowing
>>>stuff
>>>
>>>* lacking query capability
>>>
>>>* tile caching
>>>
>>>* probably several other things I can't think of right now
>>>
>>>I think that there are some differences in how OpenLayers and ka-Map
>>>approach things, but I'm not entirely sure that they are significant.
>>>
>>>At this point, I am proposing to join the two projects into a single
>>>project. My desired approach would be to start with the OpenLayers
>>>code and start to bring over the parts of ka-Map that I(we) feel are
>>>missing.
>>>
>>>Please let me know what you think of this. And point out all the
>>>problems you can think of.
>>>
>>>Cheers
>>>
>>>Paul
>>>
>>>+-----------------------------------------------------------------+
>>>|Paul Spencer pspencer at dmsolutions.ca |
>>>+-----------------------------------------------------------------+
>>>|Applications & Software Development |
>>>|DM Solutions Group Inc http://www.dmsolutions.ca/|
>>>+-----------------------------------------------------------------+
>>>
>>>
>>>
>>>
>>>_______________________________________________
>>>ka-Map-users mailing list
>>>ka-Map-users at lists.maptools.org
>>>http://lists.maptools.org/mailman/listinfo/ka-map-users
>>>
>>
>>+-----------------------------------------------------------------+
>>|Paul Spencer pspencer at dmsolutions.ca |
>>+-----------------------------------------------------------------+
>>|Applications & Software Development |
>>|DM Solutions Group Inc http://www.dmsolutions.ca/|
>>+-----------------------------------------------------------------+
>>
>>
>>
>>
>
>
> +-----------------------------------------------------------------+
> |Paul Spencer pspencer at dmsolutions.ca |
> +-----------------------------------------------------------------+
> |Applications & Software Development |
> |DM Solutions Group Inc http://www.dmsolutions.ca/|
> +-----------------------------------------------------------------+
>
>
>
>
>
> _______________________________________________
> ka-Map-users mailing list
> ka-Map-users at lists.maptools.org
> http://lists.maptools.org/mailman/listinfo/ka-map-users
>
--
Daniel Morissette
http://www.mapgears.com/
More information about the ka-Map-users
mailing list