[OSGeo-Standards] Re: WMS-C and handling of tiled=true? Indicating a Restriction or an Ability

[Satoru Takagi]

Hi Jody,

 >I think we all are like that  :-)
It is nice! Furthermore, I have a browser-centric sense if I say. In 
here, this might be accepted, too.


GWC seems to be tiled map dedicated server which delivers the data 
collected from the server which have WMS interface.

On the other hand, I think that it cannot consider WMS-C to be a 
specification for such tiled map dedicated server.

It looks like the specification of metadata that offers hint information 
for the client such as browsers and GWC to acquire the tile data from 
the WMS server.

That is, I think that WMS-C requires all WMS functionalities.


Therefore, I thought that it was undesirable to consider WMS-C as a 
general specification for the tiled map dedicated server.

In a general specification for the tiled map dedicated server, I think 
that higher generality where the very simple architecture that I 
previously presented is included is necessary.
Therefore, a further examination might be necessary, for an enough 
examination has not been performed yet.

Regards,

Satoru Takagi


Jody Garnett さんは書きました:
> Satoru Takagi wrote:
>> Hi Jody,
>>
>> First of all, I had acted in the world different from the GIS culture 
>> that centered on OGC. 
> I think we all are like that :-)
>> If anything, I belong to the world of SVG and Semantic Web of W3C. 
>> Moreover, I am in Japan. Therefore, I might have a sense different 
>> from people of the GIS community. And, I am continuing the design of 
>> the Web Mapping Platform architecture based on SVG.
>>
> With respect to making a "TileService" that is very simple and direct 
> (and desribes a static set of tiles) - this is a great idea and one I 
> believe is being worked on. My own introduction to the topic comes from 
> the WorldWind tile servers which also make use of a nice number 
> convention where row/col can be used to retrieve a tile.
> 
> 
> In the meeting whereWMS-C was made we strongly considered a getTile( 
> row, col ) operation - at the time it was rejected in order to focus on 
> adding caching to existing WMS servers. If I read Mark's email it sounds 
> like the OGC is returning to this topic (and point of view).
> 
> But for today I need WMS-C to work; and I have to answer the question on 
> how to notice if a WMS-C is general purpose or not.  I can either keep 
> talking on this email list and hope that the implementors are taking part?
> (are they?)
> 
> Or hunt down GWC and TileServer implementors and see if we can get this 
> one sorted out.
> 
> For reference here is what I will go shopping with:
> - if the VendorSpecificParameters includes "tiled=true" as part of the 
> GetMap preamble - I will consider it mandatory - and the resulting WMSC 
> to be restricted to only tiled requests
> - if it is not included - I will add it when I make the requests - but I 
> will feel free to make any general purpose GetMap request
> 
> Jody
> 

-- 
--
高木　悟（Satoru Takagi)
ＫＤＤＩ株式会社 コア技術統括本部 ネットワーク技術本部
技術戦略部 企画調査Ｇ
Email:sa-takagi at kddi.com

この電子メール及び添付書類は名宛人のための秘密情報を含んで
います。名宛人以外の方が受信された場合は、お手数をお掛けい
たしますが、破棄をお願いいたします。