[Live-demo] Definition of "stable" software

Cameron Shorter cameron.shorter at gmail.com
Sun Jan 31 12:01:16 PST 2010

You raise some good points about "What should  we consider as stable 
software" for the LiveDVD.

I think that this question is equally valid for the OSGeo incubation 
process too, and should be added to the incubation list of questions.

I think we need to define "stable software" so that we have a clear yard 
stick and guidelines for assessing whether to put a project in our 
"stable" or "beta" section of our LiveDVD.

Here is a definition:
"The term /stable/ refers to a version of software that is substantially 
identical to a version that has been through enough real-world testing 
to reasonably assume there are no significant problems, or at least that 
any problems are known and documented."

So I suggest we ask:
* Does the project have a build process which puts our candidate 
releases which are tested before releasing a stable release?

* Does the project have a test suite which is run before putting out a 
stable release? Ideally tests should be run before putting out any release.

* Does the project have a diverse community who test candidate releases 
before putting out a stable release?

These questions should be included in the Incubation Project Status 

They should also be added to the LiveDVD build process.

Do you mind if I create a new draft of the Project Status Template, 
maybe at:

Hamish wrote:
> Cameron wrote:
>> How stable are these packages?
> The best I can provide is a subjective answer- different people
> and projects will have totally different thresholds and ideas of
> what that actually means. I spend most of my time with Debian
> and GRASS, both of which are extrememly conservative with what
> they call "stable", so that's my bias. (Debian's "unstable" is
> probably more mature than most other distros stable releases,
> so I'm not really sure how to compare ...)
> I'll try though:
> -Will they work without crashing? Yes.
> -GMT: very mature and well established for years.
> -Octave: mature and well established for years.
> -Octave mapping toolbox: I've little experience with it but it's
> an official debian/ubuntu package fwtw.
> -OpenCPN: a bit of a newcomer compared to the others but largely
> put together from established mature libraries (gdal+libS52+...),
> by an active dev team. At version "1.3.6" with packaged versions
> available for Mac,Win,and Ubuntu, for whatever that means.
> -OSM tools: I've little experience with them myself, but as they
> are standed packages used by hundreds of OSMers all over the
> world I assume they are pretty mature.
> So I'd say that all are well in the "stable" category and have
> reached the point of critical mass, but within that you can
> probably open any geophysical journal from the last 10 years and
> find a plot made with GMT, while at the other end of the
> spectrum OpenCPN only began around 5-6 years ago and has only
> applied to be an official debian/ubuntu package in the last 6
> months.
> Hamish

Cameron Shorter
Geospatial Systems Architect
Tel: +61 (0)2 8570 5050
Mob: +61 (0)419 142 254

Think Globally, Fix Locally
Geospatial Solutions enhanced with Open Standards and Open Source

More information about the Osgeolive mailing list