[postgis-users] A bit off topic, but FOSS GIS clients...

[Tim Sutton]

Hi (anonymous person :-)

See below for some comments...


>This is off-topic. But since many use
>PostGIS/PostgreSQL as a spatial database backend, I
>thought people here may be best equipped to comment. I
>couldn't be the only one wondering about this.

Quantum GIS is painfully slow rendering and searching
through data in moderately sized, and evidently even
tiny, shapefiles. That's *with* scale-dependent
display set to reasonable values.

Did you file a bug report and provide a test data set that replicates
the issue? Open source software places more responsibility on the user
than commercial software - you are expected
to contribute your experiences back in a meaningful way (if you truly
want to embrace the community approach to software development
anyway...).

>Seems QGIS is a decade or more behind even ArcView 3.2
>(still a great product after all these years) with
>regard to performance and basic (non-OGC and web)
>functionality.

Mapserver has a big head start on QGIS, as does ArcView. When I used
ArcView 1.0 it had less features than QGIS has (which is not even at a
1.0 release yet). It seems a bit unreasonable to expect Desktop GIS to
suddenly emerge "out of the ether" with full feature parity to
commercial products or open source projects (such as your web mapping
example) that have been around a lot longer. I cant speak for other
Desktop GIS projects, but QGIS is developed only on a volunteer basis
with no full time programmers / commercial backing whatsoever. We are
trying to bring you a great desktop experience, but be realistic in
your expectations, its going to take time to be at the point where we
can get QGIS to the point of ESRI ArcMap etc. When I started with
desktop GIS on linux, the only alternative was GRASS. It took me ages
just to figure out how to create a mapset and import some shp data and
display it. Now there are a number of tools that let me easily open
and display GIS data in a pointy clicky way...

>Will importing the shapefiles into PostgreSQL solve
>the data access speed issues? Is the rendering engine
>itself problematic, or is the slowness a function of
>its inability to work efficiently with shapefiles?

>I find QGIS simply unusable when working with
>shapefiles. Considering many GIS novices still work
>with shapefiles, I'm guessing the lack of
>an efficient FOSS GIS client will stall wider adoption
>of FOSS GIS.

Many, many users find QGIS quite useable for everyday browsing of
shapfile data....

>Am I mistaken, or is FOSS GIS weakest when it comes to
>the non-web based GIS client arena?

Yes, given that 4 or 5 years ago there were no Desktop GIS projects
out there at all (other than GRASS which was here when the world
began). As I said above, give it time  it will get there...

>I seriously want to like QGIS, but am having a tough
>time of that presently. Are there other FOSS GIS
>clients that can access, search through, and render
>shapefile data better than QGIS? I'm presuming,
>perhaps falsely, that uDig is no better in
>non-RDBMS-based spatial data access/render
>performance.

There are many other FOSS desktop GIS clients - did you try *google*
for this? In my (completely unbiased)  opinion QGIS is the best but if
you are a serious GIS user entering the FOSS GIS space, it would seem
logical to me that you download the various efforts and try them out
to form your own opinion about which is best for you...after all its
free software so there is little to be lost in doing a decent survey
of the available offerings....

>Finally, have the QGIS, uDig, and other folks
>considered joining forces to create a killer GIS
>client?

This is the 'cant we take all the Free Software and put it in a big
pot and make *super-software*' misconception. QGIS (C++) and uDig
(Java) are implemented in different programming languages and there is
probably ZERO chance of them ever becoming a single project (or their
programmers suddenly all jumping ship from one to the other).

>I find it depressing to see many different
>fiefdoms in the FOSS community generally.

Actually this is an essential part to open source philosophy. See for
example a recent discussion by Linux Torvalds:
http://kerneltrap.org/node/11

Imagine the time that would be lost if there was a single effort
towards desktop GIS and the project failed (through poor design,
language deprecation etc). It would be a giant set back as its
replacement started from scratch to replicate and surpass those
features. Having many open source GIS projects on the go will lead to,
by process of evolution a few good ones 'bubbling to the surface', and
perhaps one or two others nipping at its heels to become the defacto
'best' project in its place. This competition is healthy and having
different projects allows different ideas and approaches to be
explored in a relatively low risk way.

>If several
>projects merged, it could lead to one heck of a FOSS
>software product rather than, perhaps, several
>marginal ones.

The marginal ones will eventually become good and then great, or they
will dissappear....and new marginal ones will appear behind them.



>One of the beauties of FOSS is that anyone with a
>vision can start a project and attempt to create
>something better than already exists. However, that
>vision may be realized, if ever, at a glacial pace. I
>myself am impatient, and am not a software developer.
>But if I was a developer, I would want to find the
>best FOSS GIS client out there and focus efforts on
>it.

Ah the 'Im not a developer and therefore its ok to stand by the
sidelines and bemoan my fate argument'. Actually non developers have a
lot to offer open source projects that can take the load off
developers backs - for example in QGIS I spend more and more time
dealing with users, user accounts, documentation, web sites, random
requests sent to me etc. Yet when I send out requests for help to the
community to e.g. work on the documentation, it generally gets little
or no response. If you really want to do something to help get
involved and do something - most projects dont care if you are a
developer or not, only if you are a net time sink, or a net time
asset. Really just get involved in a project you care about, make a
genuine offer to help and the project will move faster forward for
your help....

>Seems to me if people joined forces more often and
>consolidated projects, QGIS, for instance, might not
>still be choking trying to access, search, and render
>moderately sized shapefiles after 5 years of
>development.

Actually (speaking for QGIS) we are really open to working with other
projects where it makes sense. Just look under the hood - we use
GDAL/OGR for data access, GEOS for geometry operations, Qt for our gui
toolkit, GRASS to provide analytical tools, and at some stage in the
future we plan to integrate OSSIM & ossimPlanet functionality. Perhaps
in 5 more years time we will even be able to open shapefiles....

>So what gives? At conferences like FOSS4G, is there
>ever talk of project consolidation? If not, why not? I
>tend to think of all the development hours spent on,
>say, 8 FOSS GIS clients, wasted, when, if
>there was focus, 1 or 2 FOSS GIS clients could really
>kick some butt and give commercial products real
>competition.

Ok I think I addressed this above already...


>Why doesn't project consolidation happen often--or not
>often enough? Hurt feelings? Unwillingness to judge
>one product over another?

...and this...

>Have there been no systematic attempts by the
>community to seriously assess what projects are out
>there, find 1 or 2 best of breeds, then encourage the
>focus of development on those?



>I appreciate all the work that's been done on QGIS,
>uDig, and others. But I personally would love to see
>more consolidation so we make larger, quicker strides.

... this I addressed above too I believe.... want to really help?
Sponsor me (or others like) to work on QGIS / uDig etc full time so we
dont have to waste all our precious hacking time on these pesky day
jobs.... or start pitching in to the project you care about to make
things run more smoothly...or do *something* to make the project you
care about succeed....


>Final question--if I import large shapefiles into
>PostgreSQL/PostGIS, and use QGIS or uDig, will my
>speed  and usability gripes be extinguished? Honestly
>can't ever imagine using the latest QGIS with
>shapefiles for more than 10 minutes without wanting
>immediately to uninstall it.

File a bug & work with developers to fix your issue - turn your
problems into an actual contribution and the projects will be better
off for it....


As a final note I would say that the general tone of your email is
rather bleak, but in fact there are many many users who are happily
and productively using QGIS, uDig etc for day to day work. I
appreciate that the software out their may not meet all your needs
yet, but for Joe Bloggs in darkest Peru at least there is *something*
out there for doing basic Desktop GIS without needing to spend 5 years
salary on an ESRI license....

Have fun!

Regards


-- 
Tim Sutton
QGIS Project Steering Committee Member - Release  Manager
Visit http://qgis.org for a great open source GIS
openModeller Desktop Developer
Visit http://openModeller.sf.net for a great open source ecological
niche modelling tool
Home Page: http://tim.linfiniti.com
Skype: timlinux
Irc: timlinux on #qgis at freenode.net