On 10/12/2015 11:57 AM, Robert Green wrote:
"Don Y" wrote in message
...
On 10/5/2015 7:14 PM, Robert Green wrote:

stuff snipped

I was building 3D CAD models some 20+ years ago (AutoCAD v11 w/ AME).
I can recall having a problem with the package (some $3K as an *upgrade*)
and having a fix in my hands within days.

If I had a similar problem with a FOSS product *today*, it would probably
be weeks for someone to "take an interest" in my particular problem,
devote some time researching it and then days or weeks for someone
to decide it was worth *fixing*!

That's a pretty broad brush you're tarring all FOSS creators with. I don't
accept it. I've worked with a number of Home Automation software publishers
who have literally fallen all over themselves to fix a bug or issue that I
identified and do it very quickly. If they couldn't fix it right away
because of some complex dependencies elsewhere in the code, they'd certainly
file it away in their "we'll get to it at the next major build."

When's "the next major build"? Is it before or *after* my release?

E.g., it is not unusual for a database to be a read-only construct and
still be useful. In fact, it can be a *desirable* characteristic
of a database -- the CERTAINTY that the contents CAN NOT be altered!

I.e., put the "data" on read-only media. Try as the software might
(bugs, malware), there's simply NO WAY to alter the persistent copy of
the data!

I rely on this capability in my current project. I.e., the data
resides *in* read-only memory. *If* the DBMS expects to be able
to alter it FOR WHATEVER REASON, that action WILL fail! There
is simply no way to write to the memory even if you deliberately
tried to do so!

Requests for this feature/capability are simply not important enough
(apparently) to rise to the level of "active development" (for
PostgreSQL). The feature *is* apparently supported under Oracle.
And, IIRC, under MySQL (though possibly as a kludge).

Given that I've adopted the philosophy that everything I'm building must
be available under an "open" (nonGPL) license, Oracle is out of the
question. As I've not been impressed with MySQL, that leaves me
with the only choice of taking ownership of a PostgreSQL release and
adding the features that *I* want *to* that release -- taking full
advantage of the license terms to do so as a "spin-off" codebase.

Of course, my priorities aren't the same as the PostgreSQL development
team -- nor its user base. So, my modifications will be of little
direct use to them. *But*, they'll meet *my* needs. If, at some
future date, someone wants to backport new PostgreSQL features to
my implementation; or, port my features to -CURRENT, that's entirely
up to them to do so -- without it impacting *my* efforts, time table,
etc.

Interestingly enough, when I asked for features that would benefit mostly me
and not all the other users, I always got the "we'll see" which is
parent-speak for a deferred "no." "Can I have a BB gun for Christmas?"
"We'll see."

Meanwhile, the clock is ticking on *my* project. What do I tell *my*
client? "The FREE software that I'm using doesn't work correctly.
I'm hoping someone will step up and offer me a solution sometime soon.
I realize *you* have market constraints that are pressuring you for a
product offering but there's nothing *I* can do to speed things up..."

If you told me that, as the guy who hired you, I would simply say "It's
unfortunate that you used tools that were unsuitable for the task. Didn't
you evaluate them fully before starting?" Whose fault is it then? The poor
FOSS creator who's doing it as a labor of love or the guy trying to save a
buck by using FOSS on a paying project?

Exactly. I spend $15-70K/year on tools. Because I *don't* want to ever be
wondering why a *tool* doesn't do what I expect it to do. Saving a few
dollars (or, even a few tens of kilodollar) doesn't make sense when my
reputation and a client's *product* (plus his reputation) are on the line.

I did projects 30 years ago that FOSS software *still* isn't up to the
task to address! Let alone do so efficiently and with minimal effort.

I rarely bring freeware onto a paying project but if I do, I make sure it's
functioning and that it can be properly licensed by the end user.

I've never had an issue with IrfanView, VLC, HexEdit, WinZip, PKModem (in
the very old days) and so many more products that I have just the opposite
view of FOSS than you have. What FOSS program burned you so badly that
you're willing to classify them all as drek?