[OSGeoJapan-discuss] Integration of Japanese GIS formats into mainline GDAL ?

[Hiroshi Miura(@osmf)]

Hi Minoru,

Thank you for your primary effort in HackFest.
 zero to one ;(c) paypal mafia :)  is an one of  greatest things
and I'd like to express appreciation.
ありがとうございます。

On 2017年03月18日 11:26, Minoru Akagi wrote:
> Hi Miura-san, Even and all,
>
> The FGD schema has been updated many times. Frankly, GML is too
> complex. It is extremely difficult to create a software that fully
> supports all the files witten in the language. We need to avoid
> breaking the GML driver compatibiilty with the data formats of other
> countries. Tinkered driver code will burden developers and maintainers
> (sorry! :-<).
It is because we need more test data and test cases  not to break other work.
GDAL is a collaborative work and is a platform to help each other over countries.

Please don't shame but let's gather your and Japanese community
to contribute to the code.  Tinkered code can be forged
like Japanese sword artisans team by pushing, pulling, hardening and patching:)

Ok, it is true  GML spec is very complex and GSI's crafted specification is not stable.
There is a good chance to put our ideas to separate these complexity from the code.

GMLAS is an approach toward this idea.

I've just PR to OSGeo/GDAL then I can get TravisCI/AppVoyer automated test
process and got response from anger Jenkins ;-P

This process helps us from a fear.

ですので、ぜひテストケースを沢山準備しましょう。
他の国のGMLのテストケースも沢山準備していけば、
壊す心配をしなくていいでしょう。
> I like Miura-san's idea to convert FGD files with a Python script and
> GDAL library. Python script converter is easy to be updated and can be
> released frequently. It must be a good way for volunteer developers.
>
Same things are on the script. It it not support all the data.
Any difference? Duplicated and separated effort is worst selection for us, OSS community.

同じ事です。大きなGDALというソフトウエアを壊す心配して、
分かれた努力をするより、一緒に取り組んだ方がいいです。

自分のためにオレオレLinuxを開発して、廃れていった
事例はいくらでもありますよねぇ。

Hiroshi