|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | Re: GISList: RE: OGC and Standards, - a response |
| Date: |
12/11/2002 08:17:31 PM |
| From: |
Frank Warmerdam |
|
|
Anthony Quartararo wrote:
> "..You are quite correct that some of our members are the traditional
> GIS vendors. However, probably 80% are not!.."
>
> I'd wager lunch that the 80/20 rule applies here.
Anthony,
I thought it might be helpful to have a view from a little further
outside OGC. From my involvement with OGC over the last three years
it does not seem dominated by the big GIS vendors to me. From what I
have seen, much of the (recent) specification work has been done by
relatively small GIS companies such as Cadcorp, Cubewerx, and Galdos.
I think it is very wrong to paint OGC as being a vehicle for the big
GIS vendors to run the industry.
> "Moreover, all approved OGC specifications are freely available to ALL
> developers of GIS and other software, worldwide. That means that even if
> a software development firm never pays a cent to OGC, it can implement
> any OGC specification. There are no associated royalty fees. This free
> and open, unrestricted access to OGC specifications is especially
> valuable to both commercial software providers as well open source
> software providers, several of whom have implemented OpenGIS
> Specifications in their offerings. "
>
> So what is the compelling reason to become an OGC member ? How many open
> source (GIS) software providers are there ?
I am not sure what your point about "how many open source software
providers are there?" is. The point seems obvious, but OGC and
OpenGIS is not about open source. However, as a developer of open
source GIS software I think the OGC work is very useful, and gives me
a greater opportunity to provide software components and interoperability
technology to my user base without just being tied down to one proprietary
vendors architecture.
I don't know that you have a compelling reason to become an OGC member.
Other than hearing what you don't like, I am not sure what you do like
or want.
One thing I will suggest is that I think it is important for more user
organizations to take a driving role at OGC. As far as driving the
agenda at OGC, I think a variety of American federal agencies (military
and civilian) have had the largest role, and acted on behalf of the
user community as they see it. And I think that is very helpful, as
opposed to leaving things totally in the hands of the software vendors
themselves.
> "As a matter of fact, the larger GIS players are to be commended for
> their willingness and commitment to open their systems by implementing
> OpenGIS specifications."
>
> Providing a method for one proprietary system to talk to another
> proprietary system is different than "opening" your system up. This is
> nothing new. Many consultants make a nice living providing enterprise
> application integration services, which essentially does the same thing.
Providing standards based specifications for data models, formats and
software interfaces is a huge step towards "opening things up". Just
what do you think should be done. If you think the only kind of "open"
is open source, come help me, I have lots of software that needs writing.
But don't denigrate other important kinds of openness.
> "As for conformance, there is a big difference between implementing
> products and conforming products. There are hundreds of implementations
> of OpenGIS specifications in the market place - these are products whose
> developers have included one or more interfaces in the product. A number
> of these are conformant, that is, have passed rigorous tests
> illustrating that the interfaces behave as indicated in the
> specification. A new Web services conformance testing framework will
> make conformance testing even easier in the future."
>
> Don't you see mass confusion with this? Conformant, compliant,
> implemented? Add these to a list of overused, ineffective "marketing"
> terms that have flooded the industry in recent years and you will find
> consumers of all types tuning out to the subtle, yet important
> differences, and if they tune out, it no longer has value.
Frankly I am thrilled that OGC has seen fit to list software packages that
I and my associates have developed to OGC specifications on their list even
though I can't afford the substantial costs to have conformance verified.
I would add that conformance testing is expensive because it is hard, not
because this is a big "profit center" for OGC. As far as I know there are
just two concepts here - "implementing" meaning untested and "conforming"
meaning tested.
> No insult intended, please accept my apology. But why not start from
> scratch? Impractical? Impossible? No money ?
Well, the obvious reason to not start from scratch is that it would set
the whole
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
