Re: Little activity ?

> From: "Arne Chr. Jorgensen" <instel@...>
> Just wondered if anyone are familiar with LinuxCAD ?
> http://www.softwareforge.com/linuxcad
>
> This is a replacement for AutoCad, under Linux. Is it any good ?
>
> Could be interesting, as I am familiar with AutoCad.

I bought a copy of this junk without doing any research first. Luckily I only
lost $99. Some relevant links are:

http://pw2.netcom.com/~rwuest/linuxcadreview.html

and

http://www.zip.com.au/~erikd/lcad.html

Over the years I have bought a lot of software and most of it has been a
disappointment. I don't mind paying for software. It's just that there's
little to no actual support to be had. If you find a bug, or need another
feature you are wasting your time if you call the manufacturer.
I would be really interested in finding a way to pool the financial resources
of a group of people who all wanted a certain software application written
(CAM software for example). We could then hammer out a spec and hire some
programmers to make it a reality. Once its basic functions work (that is, it
has enough function to be useful and not some alpha stage curiosity) and it's
RELIABLE we release it, with full source, under a loose license like the GPL.
You might ask, "Why give away the store?". The answer is that once you get a
project like this going it tends to become self perpetuating. My hope is that
other people will use the basic application as a platform to which they add
new features they need. Under the GPL these improvements get plowed back into
the publicly available version and the application grows. The trick is to
gain the necessary momentum and burst onto the scene with something good
enough to attract attention. I believe that contrary to the "release early,
release often" philosophy of some open source advocates that you need to
bring something substantial to the table early on. A good example of what not
to do is:

http://www.fpa-engineers.com/OD/

I'm sure these guys mean well, but I think they'll have a tough time
attracting the interest of developers without at least SOMETHING to build
from.

What we could do is what Microsoft does. They write some of their own stuff,
but to save time they also buy whole companies to acquire technologies they
need rather than reinventing the wheel (they also do this to eliminate
competition, but that is another story).

WARNING - THE FOLLOWING IS ENTIRELY HYPOTHETICAL!!!

Ron Ginger has written a really nice little CNC/CAM program (I mean it really
is nice, that's not the hypothetical part). It's not completely finished, but
there's a lot there including some really nice graphical toolpath displays
and the ability to generate CNC code for a MillPower system and perhaps G
codes without much more work. He has written a few posts describing his
efforts and soliciting advice on what direction to take in his development.
He has indicated that he might turn it into a commercial product when it
achieves its final form, but the full scope of the program's functionality
hasn't yet been specified (that's the nature of the advice I believe he was
seeking).
Suppose we could come up with enough incentive (probably in the form of
money) to compensate the author for his work so far, and also to continue the
development enough to get the program "up on its feet" in a form that's
useful to those who provide the incentive. From my own standpoint there are a
few drawbacks to this program, but they are probably not insurmountable.
First, due to my interest in the EMC software, I'm not really concerned with
the CNC part of the program except to the extent that algorithms or
techniques in the code could be folded into the EMC code in order to improve
it. Second, it needs to be translated to a multiplatform environment,
probably Tcl/Tk (not Java, been there, done that, too slow). This would allow
it to be combined with Dave's Tcl/Tk gui for the EMC at:

http://users.nni.com/daveland/metal.htm

and also the Tcl/Tk stuff Fred Proctor's been working on lately. This would
give the EMC the "conversational programming" feature it needs as well as
allow folks to program offline on Linux or Windows if they prefer to sit in
the office. It might also be possible to pull in other resources such as the
G code generating routines from Jon Elson, and code from the GnuCAD project.
I don't know enough about software to be able to judge whether merging these
projects and converting to Tcl/Tk is more or less economical than starting
from scratch, but I think you can get an idea of what I have in mind. I know
there are huge holes in what I've proposed, such as how to coordinate the
efforts of multiple developers whose locations aren't the same, or who gets
to approve of the spec, etc. I present this as an example of how a worthwhile
goal might be attained. Well, it's 3:15am and I have veered WAY off the topic
of "Little Activity" so I'll stop here.

Thanks,

Matt