[OSGeo-Discuss] OSGeo Next five years

Chris Puttick chris.puttick at thehumanjourney.net
Wed Sep 16 23:29:59 PDT 2009


Not entirely an OSGeo specific point, but this type of criticism "open source is hard, closed source is easy" is not uncommon. Some 17 years of experience working with IT in organisations gets me a one word response to such a claim I will not repeat here, but it starts with a "b" and sounds a lot like hillocks...

What closed source, marketing driven, products tend to focus on is the appearance of easy. It has been easy to set up a Microsoft Exchange server for over 10 years, in the sense of starting with a server and successfully sending and receiving email, etc.. As so many compromised servers over the same 10 years easily demonstrate, it is hard (and requires expertise) to setup said server to only send/receive email just for those who should be able to send/receive email, hard to get said mail server to scale with your organisation, hard to unpick it after a malware infestation (or bad AV update), hard to migrate it to another mail server, nearly impossible to use it from your choice of desktop platforms and actually impossible to deploy it on your choice of server platforms. There are many other examples, many of which from Microsoft, with the same story - deceptively easy to get going, desperately hard to make it do what you finally realise is best for *your* organisation.

Simplicity in the sense of does not require expertise to make work almost certainly means impossible or very, very difficult to finally make it work the way you need it. My exposure to .net developments and the final convoluted efforts developers go through to bend to their will have provided sufficient evidence for me to tell colleagues in other organisations that it is a mistake to be deceived by rapid early progress. The tortoise and the hare is a very old story...

So if you want simplicity, put some of the building blocks together for the new user; build VMs with complete working setups that just need network configuration and data to start doing things. But please don't repeat the approach of the marketeers, make something simple and restricted and then claim something that just ain't true!

Cheers

Chris


----- "Arnie Shore" <shoreas at gmail.com> wrote:

> As a very interested lurker, and as one who has developed an Open
> Source Computer-Aided-Dispatch system that has embedded google's maps
> product, I can tell you that one of the deterrents I see is the
> relative complexity of an Open Source GIS implementation - as compared
> to the use of GMaps, which also, of course and notably, is free. The
> single source of both the tiles as well as the API is relatively
> straightforward for the non-cartographer novice.
> 
> My user community includes a fair-sized portion who have never before
> implemented a web-server-based system, and our package is designed to
> minimize the number of elements that need separate collection and
> configuration. To tell them that they need a map server in addition to
> the stack that WAMP, XAMPP, MAMP, installs in a single executable will
> turn away too many candidates, IMO. In our case, the tile-serving
> capabilities could be met by a rather limited set of server-side
> functions that are OL-aware. But I haven't seen anything like that in
> the panoply of products that comprises the OSGeo world. Please correct
> me on this if such exits.
> 
> (Further evidence of the importance of the ease-of-implementation
> issue is the proliferation of open source libraries that include
> capabilities taht are based on a GMaps foundation.)
> 
> I will say that my users - many of whom are into emergency operations
> - indeed are asking for an implementation that wd allow operation
> while disconnected from the Internet. Impossible in a GMaps-based
> solution, but completely feasible in one based on OpenLayers plus
> locally stored OSM tiles. Users I've pointed to the available OSM
> sites have told me that the level of detail wd be completely
> satisfactory as a suitable replacement for GMaps. Which is a
> critically important data point, IMO.
> 
> My perception of the current evolution of the world of Open Source GIS
> is toward greater complexity and richness. Which certainly makes for
> excitement and challenge for its enthusiasts; but that isn't doing
> much for those of us along the borders looking over the fences, and
> with limited hours available to hop that fence and get involved.
> 
> Make entry easier than it is, folks. Please?
> 
> A. Shore
> Annapolis, MD
> 
> 
> 
> On Wed, Sep 16, 2009 at 5:09 PM, Ravi < ravivundavalli at yahoo.com >
> wrote:
> 
> 
> Hi,
> have been going through all the wishes, all the arguments about how
> Open Source GIS must evolve etc. ...
> 
> _______________________________________________
> Discuss mailing list
> Discuss at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/discuss


------
Files attached to this email may be in ISO 26300 format (OASIS Open Document Format). If you have difficulty opening them, please visit http://iso26300.info for more information.




More information about the Discuss mailing list