This article http://www.linux-watch.com/news/NS7891815881.html A
medical open-source legal hell-hole Dec. 27, 2006 makes reference to an
Open Source MUMPS. Does that mean there is in essence an Open Source
Caché?

Henry Keultjes

HI Henry
You really have turned one up here. One will watch with interest - the
destruction of the English language by the combatants solicitors is
intriguing on its own, never mind the reverberations for Open Source.
Happy New Year
Peter McMurray

<hbkeultjes (AT) gmail (DOT) com> wrote

This article http://www.linux-watch.com/news/NS7891815881.html A
medical open-source legal hell-hole Dec. 27, 2006 makes reference to an
Open Source MUMPS. Does that mean there is in essence an Open Source
Caché?

Henry Keultjes

Doing a Google search on: linux gt.m

returned the following:

http://www.sanchez-gtm.com/

The reference to MUMPS on the page certainly makes this look like the
product referred to in Henry's article.

Cheers,

Brian

hbkeultjes (AT) gmail (DOT) com wrote:

Henry wrote:
Relating GT.M to Cache' is like suggesting that since OpenQM is OSS
that Universe and jBASE should or might be as well. It's
non-sequiter. D3 and mvEnterprise have common roots going back to
R83, as Cache' and GT.M have a common code base in MUMPS. The
products have a common history, not destiny.

Cache' is the combination of a few flavors of MUMPS. It's fully
commercial, and forked off of anything that came prior. One of the
flavors, not acquired by InterSystems was GT.M. That was owned by
other companies and their implementation was independently released to
GPL. That doesn't affect what InterSystems does with their
implementation.

MVCache' has nothing to do with this GT.M thing or the battle taking
place with this VistA software. And as MVCache is yet just another
word in the book of their family history, this stuff is completely
removed from our community.

And Henry, let's put this in perspective. InterSystems has done an
amazing job of marketing Cache' to new audiences at a very reasonable
price. Open Source doesn't buy this software or company anything more
than they already have. Open Source is a strategy toward a goal. If
you achieve the goal through a different strategy then alternatives
mean nothing. Cache' might be open sourced at some point, I don't
think anyone's crystal ball is big enough to see that one, but for our
purposes of writing business software which can be extended to broader
markets, even that initiative would mean nothing to this community who
doesn't make use of the open source products that are already
available. Open Source for the sake of open source is a questionable
solution in search of a problem.

T

hbkeultjes (AT) gmail (DOT) com wrote:
Hi Henry --

Nope. Cache's is in the family tree of MUMPS like UniData is in the
family tree of PICK. OpenQM is in the PICK family tree too, but that
doesn't mean there is an open source version of UniData. Cache' is
also more advanced over the open source MUMPS than UniData over OpenQM
as best I can tell.

By the way, Bohner mentioned your comdex chairs with footpeddles today
(I hope I didn't confuse two stories, I might have to check notes
rather than relying on my brain). I like the idea of the control key
in a foot peddle. Given the peddles used with pianos and sewing
machines have lasted a so long, it is surprising that there is no
footpeddle used with computers. I like the idea. Cheers!

--dawn

Tony Gravagno wrote:
That's exactly my point though. Because of the unique portability of
applications betwen the various versions of Pick, one Open Source
version gives Open Source devetees the opportunity to participate in
Pick.

Henry

Henry wrote:
Yes and no for a few reasons.

Pick is not "uniquely" portable. Relational databases are portable
too. Migrations are no more or less difficult than similar migrations
across relational platforms. We have lots of things to brag about but
this aint one of them. If Pick were as portable as people pretend to
think it is then more of the third-party products in our market would
be cross-platform. As it is, most vendors tend to focus on specific
platforms, with some platforms supported in part or not at all,
because maintaining their software for more environments strains
resources beyond profitability. We used to be more portable back when
licensees were R83 compliant and SMA had some influence on
development. But all of the licensees are gone now and all of the MV
variants develop completely independently, adding some compatibility
only as a lure for one-shot migrations, not as a way to facilitate
real cross-platform development.

OpenQM is not for use in production environments. The open source
nature of it means absolutely nothing in terms of real world use. And
since the open source version of QM is a couple months behind the
commercial, for fee product, one can't tweak the source in the
production software. The process there is to go back to an old
version, submit your changes, and hope that Ladybridge approves of it
for some future version. (And by the way, your "open" source now
belongs to Ladybridge whether they use it or not.) Most people who
are really into open source expect their _current_ software to be open
source, not just the two month old developer version. Of course there
are exceptions - there are purists who are just comfortable with the
idea that the source is theirs, should the authors take a bus into the
twilight zone. Then again for commercial software we can put source
code into escrow to achieve the same ends, so again open source
doesn't buy people a lot here.

As far as giving people an opportunity to learn the MV environment,
source code doesn't help the DBA who actually needs to use the DBMS.
Good documentation is what will endear people to this or any software,
not source code in C. If you're confusing free beer with open source
as most people do then remember that almost all of the other MV
vendors provide free licenses for developers, and some provide free
licenses for educational or non-profit use. Everyone has ample
opportunity to learn the MV model without any open source.

Consider the audience we're talking about.
The Pick model was originally adopted largely by non-IT people
who found that they could manipulate their computer system and write
their own software without a lot of education or assistance. This was
back in the day when computers were mainframes and having one on the
desk was a dream. Sure I'm painting with a broad stroke here but if I
ask for a show of hands I'm sure we'll see a good number of them.
Over time people learn the craft of IT using MV as their guide. Pick
facilitates learning IT after one has already been exposed to the
environment.
Open Source is for C/C++ developers who come at each new
package already with a great deal of knowledge about how software
works. They don't need open source to introduce them to a new data
model. They aren't going to approach the data model from the inside.

What we still need in this market is real documentation, tutorials,
source code samples for how applications (not system utilities) are
constructed. With that sort of basic information newcomers can learn
the environment, compare it with others, and come to their own
conclusions. You're confusing the open source nature of software with
the ability for people to learn how the software is used in "real
life", as if looking at the source is going to endear them to the
model. That's not the way we learned it - we like Pick in part
because it's easy to write BASIC code and command-line queries.
Jumbling someone's head with C code isn't going to give them that same
perspective.

There are key stumbling blocks to introducing these products to
newcomers. None of this is being fixed by open source. This market
has been around for 30 years and product installation is often a pain.
We still have very little documentation for what to do when you get to
that command line. We have poor or non-existent sysadmin tools. Some
people are looking at open source as though it's somehow going to be a
saviour for the model. I don't see many people contributing to public
documentation projects or open source sysadmin tools, which is where
we're going to get a lot more benefit than open source DBMS code.

Again I'll mention open source GT.M isn't related to Cache', and
therefore none of this translates to MVCache' or any of us. If you
want to talk about open source, fine, but I still don't see a relation
between the OSS in the OP and our market.

Regards,
T

Note1: I've opened a number of subtopics and have not gone deep into
any because we've sort of beat these horses dead in the past. As
written, people can argue with any of the statements above - I didn't
intend to make any complete arguments and I won't once again respond
in detail when someone points out the obvious.

Note2: Martin has done a great job with QM/OpenQM and the whole OSS
initiative, I have no problem with any of that. Anyone who implies
otherwise is a shmuck. For reference, QM (over Windows, commercial)
is probably the easiest software in our market to install and use
immediately, and the documentation is very good. Though neither the
installation nor the docs are open source, so I don't think we can
point to these as examples of the benefits of the OSS model.

pedal, unless you're selling feet.

happy new year

chandru murthi

"dawn" <dawnwolthuis (AT) gmail (DOT) com> wrote

hbkeultjes (AT) gmail (DOT) com wrote:
Hi Henry --

Nope. Cache's is in the family tree of MUMPS like UniData is in the
family tree of PICK. OpenQM is in the PICK family tree too, but that
doesn't mean there is an open source version of UniData. Cache' is
also more advanced over the open source MUMPS than UniData over OpenQM
as best I can tell.

By the way, Bohner mentioned your comdex chairs with footpeddles today
(I hope I didn't confuse two stories, I might have to check notes
rather than relying on my brain). I like the idea of the control key
in a foot peddle. Given the peddles used with pianos and sewing
machines have lasted a so long, it is surprising that there is no
footpeddle used with computers. I like the idea. Cheers!

--dawn