Subscription questions...

[Fairlight]

On Tue, Mar 15, 2016 at 05:03:21PM -0400, Richard Kreiss thus spoke:
> 
> If memory serves, fpTech announced (Bud) that they were freezing filePro
> development between 5.0 and 5.6 release to correct as many of the bugs as
> they could to create a "stable platform".  This meant that there were no
> interim bug fixes after 5.0.15 which had the license manager.  Therefore
> when 5.6 was rolled out, it contained most or all of the bug fixes and had
> some new features. One complaint from many of you was that it took too long
> for new releases to come out.  

And they promptly buried those fixes in a version which many don't want to
upgrade to -because- of that license manager.

> Also, I think that we got used to getting maintenance releases with free
> upgraded functionality.  To run a viable business, you can't keep giving
> away product.  This added functionality costs money to create and as a
> business one would expect to profit from these enhancements.  FpTech changed

Nobody disputes that.  The model should have been split out decades ago.

> As for maintenance releases, what is a reasonable time period for creating
> maintenance releases for an older version?  Should they still be doing
> maintenance release on 5.0 or 5.6 or maybe just 5.6 and above.  Maybe I

They should -have- done maintenance for a substantial amount of time after
the release of the new version.

5.0.15 never should have had the license manager ported in, for one.  But
getting past that, 5.0.15 should have been supported for probably five
years after 5.6 came out, and 5.6 the same after 5.7 came out, etc.

I'm thinking something around Red Hat's life cycle, which is fairly
reasonable for enterprise solutions.

Instant obsolescence the second a new version is released is not really a
good business practise.  I don't do it.  I generally wait at least a year
before I've EOLed a version of OneGate once a new version is out.

> should have asked the question "What is a reasonable length of time for the
> end of life for a particular release version?" I think most software
> companies will go back 1 release; although some may do 2.  That would be
> having to maintain the current version and 1 to 2 older versions.  This
> would be a big financial drain on any company having to keep a programming
> staff working on 2 or 3 versions of the software.  

The main reason I EOL is so I don't have to maintain two trees.  I'm also
one person, so it's a bit more burdensome to do maintain more than one
tree.  But my stuff goes years without -any- credible bug reports.

For enterprise-grade software, I'd say a decade-long life-cycle in general
is reasonable, with at least two active branches.  That's about what RH and
CentOS do.  So right now, 5.7 and 5.8 should both be getting maintenance
releases, in my opinion.  I'd have to look up the release date for 5.6 to
see what I think.

Due to the horrible way 5.0's EOL was handled and intermingled with 5.6,
that's a special case which doesn't even follow sane rules.  That was,
frankly, a nightmare version of how -not- to EOL a product.

mark->
-- 
Audio panton, cogito singularis.