|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | RE: [gislist] google maps and OGC Testing |
| Date: |
02/23/2005 07:10:01 AM |
| From: |
Jeff Harrison |
|
|
OK, I can't help myself any more. This discussion started with the topic
of Google maps, so let's deal with that first. I'm happy Google has a
web mapping service. That said, the conversation turned to the
conclusion that "well, this Google Map service and others out there make
OGC and other standards unnecessary."
Sorry, this is not true. The idea behind OGC web mapping was that a
client can use map services from multiple sources, in one multi-source
view, increasing the ability to reuse already-hosted content. So, Google
Maps highlights once again that folks can stand up a single source of
mapping content, deploy it as a map service and access it with a web
browser. But I can't integrate this content with other mapping services
without some level of interoperability. Thanks Google, now try a WMS
interface on top of the service so I can reuse it and aggregate the maps
with other sources (they might not want to do that however, since they
might want to create a "walled garden" and drive internet users to
Google Maps, whatever, that's their business).
Let's deal with the performance issue and slay the myth that OGC Web
Services are somehow not "fast enough" or that there is something in the
specifications that makes things work slowly.
Sorry, this is not true either. Folks have been spreading inaccurate
data about WMS performance for a while now. WMS works just fine, I get
maps from multiple sources in blinding speed from many vendors.
Performance factors like map generation, response time etc. are not a
factor of the specification, rather the vendor implementation of the
spec. Some vendors are better than others so if you have a slow WMS, ask
your vendor why they can't get it right. Oh by the way, not all WMS are
created equal, some folks concentrate on optimizing feature-based WMS,
others focus on imagery-based WMS, so the vendor needs to know what they
are doing. Again, the specification works just fine and is designed to
allow the aggregation of content from multiple sources.
Now let's deal with the challenge below "to anyone to make performance
metrics available for general consumption on various scenarios of
OGC-certified technology". You assert that no testing has been done.
Sorry, this is not true either. The first level of testing that has
been done over and over again is "interoperability testing", that is,
the ability for multiple services to implement a specification and work
together to satisfy the requirements of a certain set of Use Cases. This
testing has been done extensively over the last five years with the
"results" Anthony demands being shown (not just written about in a
"publication") in public forums time and time again. Here, I'll give you
one of the most recent, you can read it at -
http://www.geointelmag.com/geointelligence/article/articleDetail.jsp?id=134279
Interoperability Testing has demonstrated the network performance of the
specifications, so let's get beyond that. Oh, somebody will probably say
that WFS doesn't work and that GML is too "bulky" to "perform" well--
Sorry, that's not true either, it's works just fine for me, and it's
getting even better with compression, etc. (I can cover this iin a
separate email :)
The next level of testing is Conformance Testing. First, why don't folks
go look at -
http://cite.occamlab.com/test_engine/
and read about the Test Engine. If you like, you can test a WMS here -
http://cite.occamlab.com/test_engine/wms_1_1_1/wms_1_1_1.html
or you can also take a look at the Assertions here
http://cite.occamlab.com/test_engine/wms_1_1_1/files/wms_assertions/
Maybe I shouldn't get into the WFS conformance testing stuff (this email
is getting too long as it is) I'll leave that as homework.
OK, the next thing you'll say is, "Well, sure Jeff this is all great,
but I want to know just how 'operable' the OGC specs are not just how
'interoperable' they are so you now have to prove that to me and show me
just how fast and performant they really are". OK, I talked about the
interoperability testing above as the start of this. If the multiple
demonstrations are not enough, why don't you tell me what you want me to
test against and I'll see what I can do to get all the "metrics" required.
Or, we could just accept a simple fact. OGC specifications work. They
have been proved time and time again and they are now being used to
support geospatial enterprises.
Here's my contact info below if I can help further with this issue.
Regards,
Jeff
Jeff Harrison
mobile: 703 628 8655
creediii@mindspring.com wrote on 2/22/2005, 8:01 PM:
> So I guess the whole web services and service oriented architecture
> thrusts are invalid because they a
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
