[OSGeo-Discuss] The costs of your great productivity [was: ... is being discussed on slashdot]

[Adrian Custer]

Hey Peter,

Congratulations on all the work you have been doing. You are a powerhouse!


However, a word of caution in your work (that I have been meaning to 
raise with you for a while now).

Your speed means that you are working in a bit of a vacuum. There are 
not enough of us doing the serious analytic work at the OGC to ensure 
that all the stuff you are creating has been fully reviewed. That means 
there is little reason to think that it will be harmonizable in the 
future. Perhaps that is not of great concern right now, I don't know. 
But the less our visions align the harder harmonization will eventually 
be. So you are building rapidly a complex structure that works for you 
and your users which is fantastic. Unfortunately, there is nothing that 
is keeping us working on parallel tracks so it is likely that your 
complex structure and that of others look completely different and 
address completely different interaction mechanisms.


For instance, just yesterday I took a first look at your CRS Name Type 
Spec. and grew more concerned that I have been. I hope someday to make a 
serious review of it but have other priorities. So today, I don't even 
understand why it has that strange name. However, I *do know* that CRSs 
at the OGC are mis-designed. They work fine because essentially 
everything is tied to WGS84 but they are not semantically correctly 
structured which impacts other specs like Geometry. It does not look 
like you are even aware of this from my quick glance at the standard 
yesterday. Furthermore, the parameterized CRS stuff seems like it is 
going to need work. But maybe not, I won't know until I take a few days 
to read it all very closely to see what it is trying to do and how well 
it succeeds at that and what it does not do and ... , something that 
won't happen anytime soon. So, in working ahead so effectively, you are 
also loosing touch with others of us, or at least with me, who cannot 
track all your work. That's a tradeoff of your great productivity!



As another example, you are forging ahead on REST. Maybe. As you well 
know, everyone has a different idea what that is and, unfortunately, 
also a different idea of what the OGC should be standardizing. I can see 
that designing one pattern of use of HTTP operations and one particular 
set of URL templates is relatively easy. However, it seems to be exactly 
what we should not be doing for a long term robust, flexible spatial 
data infrastructure. My question on that front got zero response from 
the TC recently so I am not sure we are even playing the same game when 
we are thinking about REST. I guess we'll just have to live with that, 
and fix the pieces eventually. However, I do get the sense that everyone 
had decided to work in their own silos since at least there we are 
minimally productive. We find our own solutions that 'work for us' and 
avoid doing the shitty hard work to keep getting on the same page. That 
is going to cost us in the end.


The end result will probably be that the work of designing a harmonious 
suite of simple geospatial functionality handling coverages (both grid 
and feature based), processes, searches, and visualizations will end up 
being way harder than it should be and may never happen. If we were 
working harmoniously such a service would be a page long profile joining 
together a conformance class from each service. But we are not.


So keep up the great work since you are one of the great fires of 
productivity at the OGC. However, do be aware that the harmonization 
work will not vanish simply because we decide to ignore if for now.

cheers,
   ~adrian

PS This of course does not offer any way forwards, so does not give you 
any real advice which sucks. Also, my experience with 'words of caution' 
is that they are universally ignored perhaps because they are unusable. 
Nonetheless, this is something I have been thinking of mentioning to you 
for a while so now I have. cheers.





On 6/7/13 4:23 AM, Baumann, Peter wrote:
> I like the part of "not talk of OGC using 3rd person". Being one of the activists in OGC spec writing and group chairing, and also talking on behalf of my many colleagues in this, I'd like to point out that a lot of useful work is going on which goes largely unnoticed.
>
> Did you notice that since recently we can model temporal CRSs just like spatial ones? Can combine CRSs adhoc? Have a RESTful interface to WCS pretty much finished? Have 5 WCS Extensions just waiting for roll-out?
>
> Partly this was our fault, too: not enough information provided. As a consequence of "the ESRI case" we are more and more opening up discussions to the general public. For example, WCS stuff can be found here:
> http://external.opengeospatial.org/twiki_public/CoveragesDWG/WebHome
> Of course maintaining such information puts extra burden on us, but we are beginning to realize that it is instrumental in keeping contact with "the world outside OGC" (using this consciously).
>
> Conversely, we hope that there will be more feedback than has been in the past. DWG (Domain Working Group) mailing lists are open as well, so why not subscribe, follow, comment...it's up to all of us to improve on this essential communication.
>
> Just my 2 cents,
> Peter
>
>