Open sourced VMS as a business concept ?
JF Mezei
the current owner of VMS, but apparently still controlling some of the
source they contributed, isn't this a low process of open sourcing VMS to
hobbyists ?
Perhaps a plan might be for HP to keep the core kernel, and then give all
the rest to former engineers/hobbyists ? Consider long forgotten products
like TPU that might finally get some additions should they be given to
hobbyists. But with reduced engineering, the remaining engineers can only
sprou out new versions of VMS if they only touch more and more limited
areas of the kernel.
Ian Miller
There are more people working on VMS than you think.
Which parts of VMS do you think now belong to someone no longer with
HP?
LDDRIVER, DFU, What else?
Andrew
HP could OpenSource the whole of VMS rather like the OpenSolaris model.
This would allow HP to manage new OpenSource additions to the VMS
kernel but it would also allow OpenSource contributors to innovate in
user space.
As an example there is now a OpenSolaris distribution called NexentaOS
which is a GNU based user environment running on top of an OpenSolaris
kernel.
http://opensolaris.org/os/projects/#portal
Gives you a list of some of the OpenSolaris projects
regards
Andrew Harrison
Tom Linden
If you expand that somewhat to include software in general and go back to,
say, 1994, the list is pretty long.
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
pr...@prep.synonet.com
Decnet IV, several decwindows drivers, or part of, the now dead display PS,
parts of the WAN drivers, Motif, CDE, parts of pathworks, PL/1.
--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.
Bob Koehler
> With a larger-than-expected number of VMS engineers no longer working for
> the current owner of VMS, but apparently still controlling some of the
> source they contributed, isn't this a low process of open sourcing VMS to
> hobbyists ?
Still controlling? They may have copies, but the property rights in
almost all cases belong to HP, not the former employees.
There were a few parts of VMS licensed from other vendors, but lots
of those tended to go away already (remember display Postscript)?
Heck, I have a copy of the VMS 5.5-2 fiche. But I certainly don't
control the intellectual property that's represented on it, even
though it's years out of date.
dav...@montagar.com
> Perhaps a plan might be for HP to keep the core kernel, and then give all
> the rest to former engineers/hobbyists ? Consider long forgotten products
> like TPU that might finally get some additions should they be given to
> hobbyists. But with reduced engineering, the remaining engineers can only
> sprou out new versions of VMS if they only touch more and more limited
> areas of the kernel.
OpenVMS (and DEC) has a long history of open source. The DECUS Library
has been a long standing compendium of software all the way back to the
PDP days. Some things coming out of OpenVMS Engineering have been
"open sourced" (maybe not GPL, but you can get/modify the source), such
as many of Forrest Kenny's work with USB devices.
Documentation to write applications and drivers for OpenVMS has been
openly published. OpenVMS Internals has been openly published since V3
with the Kenah and Bate OpenVMS Internals and Data Structures, and
continued.
There is absolutely nothing keeping someone from getting a piece of
hardware, and writing a driver for it for OpenVMS today. Maybe there
needs to be a "recipe" for porting a Linux driver to an OpenVMS driver?
This would open up a large number of third-party hardware devices
(which are largely PCI, IDE, USB anymore). There is nothing keeping
someone from porting any number of other libraries and applications to
OpenVMS, and most of the projects at Source Forge and other places
would welcome patches back into the code base to support other
platforms (including OpenVMS). GNV and other improvements to the CRTL
have been major steps that HP and OpenVMS Engineering have done to
improve this process.
As in the Open Source movement - SOMEONE (and that generally is a
single individual and not HP, Dell, IBM, or any other corporation)
needs to take the first step, take charge, and start working on it.
That means taking action, rather than complaining about what HP does
and doesn't do with respect to Open Source. Of course, looking at the
Freeware CD, OpenSSL, CSWS, CSWB, PHP, Perl, MySQL, GNV, ilbxml2, GTK,
and more - it seems to me that there is lots of interest and activity
in OpenVMS Engineering and community regarding Open Source.
But Open Source development isn't a corporate activity - it's a
COMMUNITY activity. Don't just talk about what you think HP should do
- do something yourself. Pick something, and help port it to or
develop it for OpenVMS. Talk about your accomplishment, tell Sue, and
I'm sure she'll get it on the OpenVMS Web Page. If not, tell me or
Ken Farmer and I'm sure at least one of us will get it advertized.
Hardware is available. Hobbyist licenses (which are perfect for this)
are available. I'll help where I can - for example I can create a
forum/group on the Hobbyst page for your project and let you have full
moderator rights to it, and other coordination (and more that I'm
working on).
But enough of the "First HP needs to...". That's an excuse, not a
reason. I think that before HP is going to Open Source and give some
product to hobbyists for future development, there needs to be a core
of hobbyists that are successfully managing an OpenVMS software
development project. DEC/Compaq/HP have blessed the Hobbyist Program.
It's time to take it to the next level.
So, if you want to sign up, let me know.
Rob Brooks
> "Ian Miller" <gx...@uk2.net> writes:
>>
>> Which parts of VMS do you think now belong to someone no longer with HP?
>> LDDRIVER, DFU, What else?
>
> Decnet IV, several decwindows drivers, or part of, the now dead display PS,
> parts of the WAN drivers, Motif, CDE, parts of pathworks, PL/1.
Most of those above components (DECnet, WANDD, Motif, CDE), while at one
time *maintained but not owned* by EDS under contract with us, are now
maintained by us again.
--
Rob Brooks VMS Engineering -- Exec Group brooks!cuebid.zko.hp.com
Ian Miller
too few who act in a positive way.
There are people busy porting software to VMS now.
There are people writing VMS specific code - I've done a few utilities.
Others have done much more.
What sort of project did you have in mind?
dav...@montagar.com
There's a group working on an OpenOffice port for OpenVMS. They could
probably use some help.
With some of the discussions about databases, how about PostgreSQL on
OpenVMS? Especially if you could make it cluster aware.
Nice packaging for unixODBC would be good, especially if you could
include the ODBCConfig to streamline adding DSN's for FreeTDS, MySQL,
PostgreSQL, and Oracle. You can already build it, but sometimes having
a VMSINSTALL/PCSI kit is so much nicer.
And if you really want something challenging - Gnome or KDE on OpenVMS.
Anyone else with something on their wishlist?
hea...@aracnet.com
> Anyone else with something on their wishlist?
I'd say docbook and/or an updated TeX Distribution. Sadly I don't even
remotely have time to look at doing it myself.
Zane
Eberhard Heuser-Hofmann
Did someone ever heard what the Turin University produced:
http://h71000.www7.hp.com/news/ospp_turin.html ????
Eberhard
Dr. Dweeb
> <871wmwv...@k9.prep.synonet.com>, pr...@prep.synonet.com writes:
>> "Ian Miller" <gx...@uk2.net> writes:
>>>
>>> Which parts of VMS do you think now belong to someone no longer
>>> with HP? LDDRIVER, DFU, What else?
>>
>> Decnet IV, several decwindows drivers, or part of, the now dead
>> display PS, parts of the WAN drivers, Motif, CDE, parts of
>> pathworks, PL/1.
>
> Most of those above components (DECnet, WANDD, Motif, CDE), while at
> one time *maintained but not owned* by EDS under contract with us,
> are now maintained by us again.
Really ?!?!?! That is news. So what does EDS still have ?
Dweeb
Larry Kilgallen
> Rob Brooks wrote:
>> Most of those above components (DECnet, WANDD, Motif, CDE), while at
>> one time *maintained but not owned* by EDS under contract with us,
>> are now maintained by us again.
>
> Really ?!?!?! That is news. So what does EDS still have ?
Ian Miller
XD-ADA was jointly developed by EDS and Digital and is still supported
and developed by EDS. A port to Itanium is in progress.
They have various other products, and announced that their CORAL66
compiler is ported to Itanium VMS.
Richard Maher
> Really ?!?!?! That is news. So what does EDS still have ?
Wow! News to me too.
What about ACMS???
Cheers Richard Maher
PS. Didn't they do DECset as well? (CMS, MMS, LSE etc)
"Dr. Dweeb" <sp...@dweeb.net> wrote in message
news:458923e5$0$140$157c...@dreader2.cybercity.dk...
"Dr. Dweeb" <sp...@dweeb.net> wrote in message
news:458923e5$0$140$157c...@dreader2.cybercity.dk...
Ian Miller
JF Mezei
> ACMS is now supported by a HP team.
What does "HP team" really mean ? Just people working out of home wherever
they may be on this planet, or someone physically at ZKO ?
Back in the days of Digital, was VMS engineering pretty much a physical
team in a couple of locations ? (ZKP, Reading, Valbonnes and what other ?)
Or was it already very much into "telecommuting" with individuals spread
around the world ?
For instance, is the surfer boy on the beach in Gold Coast australia still
working for TCPIP Services ? Does Digital/HP have a physical office on the
beach or is this someone telecommuting to ZKO ?
With all the layoffs that have happened, have a number of former employees
been contracted to continue to do work on VMS on an as-needed basis ? (eg:
specific tasks instead of being paid to be in an office from 9 to 5 ?)
Phillip Helbig---remove CLOTHES to reply
> With a larger-than-expected number of VMS engineers no longer working for
> the current owner of VMS, but apparently still controlling some of the
> source they contributed,
In a LEGAL sense controlling it?
> isn't this a low process of open sourcing VMS to
> hobbyists ?
The moment VMS is open-sourced is the moment it starts quickly dying.
Ian Miller
Phillip Helbig---remove CLOTHES to reply wrote:
>
> The moment VMS is open-sourced is the moment it starts quickly dying.
Why?
Phillip Helbig---remove CLOTHES to reply
Miller" <gx...@googlemail.com> writes:
Simple. What differentiates VMS from other operating systems is the
fact that it is well planned. Without meaning any harm to Linus, the
fact that a student could write a unix kernel in his spare time says
more about the lack of sophistication of the latter than the skills of
the former. I don't see how VMS could continue in an open-source world
without a lot of coordination, and I don't see anyone with the knowledge
and resources willing to do this for free. A lot of work is going into
VMS now and this is funded by customers. Cut off the funding, cut out
the development. Many of the customers would not be willing to pay for
an open-source product.
The big myth here is that somehow VMS is not everywhere because it is
not free. Well, Windows is everywhere and it is not free. There are
free operating systems which occupy a niche, and free ones with many
users. And for hobbyists, VMS IS free. Are the masses signing up for
DECUS memberships to get hobbyist licenses? No.
Of course, DEC and successors goofed up big time in marketing, starting
with neglecting the academic market. But using this as an excuse to
jump on the open-source bandwagon is throwing the baby out with the
bathwater.
And another danger of going the open-source route is the danger of
approaching the RMS camp, where not only should all software be free but
people who write software for a living are brandmarked as morally
inferior. Read some RMS speeches; it is difficult to find anything more
hateful and inhumane in all of literature. If he wants to make his
software free, fine. I've given away software I've written. No
problem. But I think the concept of someone else telling me that my
software should be free is going too far. I don't want to have anything
to do with such hate-mongerers. If you think this is too extreme, do
some research into the philosophy, software and otherwise, of the famous
TLA-gurus. Be afraid. Be very afraid.
Of course, HP owns VMS and HP makes money from VMS so even thinking
about them open-sourcing it is idle speculation.
Paul Sture
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to
reply) wrote:
> In article <1166647616....@80g2000cwy.googlegroups.com>, "Ian
> Miller" <gx...@googlemail.com> writes:
>
> > Phillip Helbig---remove CLOTHES to reply wrote:
> > >
> > > The moment VMS is open-sourced is the moment it starts quickly dying.
> >
> > Why?
>
> Simple. What differentiates VMS from other operating systems is the
> fact that it is well planned. Without meaning any harm to Linus, the
> fact that a student could write a unix kernel in his spare time says
> more about the lack of sophistication of the latter than the skills of
> the former. I don't see how VMS could continue in an open-source world
> without a lot of coordination, and I don't see anyone with the knowledge
> and resources willing to do this for free. A lot of work is going into
> VMS now and this is funded by customers. Cut off the funding, cut out
> the development. Many of the customers would not be willing to pay for
> an open-source product.
Just to pick one aspect of such an idea, imagine what would happen to
the VMS documentation.
:-(
<snip>
>
> And another danger of going the open-source route is the danger of
> approaching the RMS camp, where not only should all software be free but
> people who write software for a living are brandmarked as morally
> inferior. Read some RMS speeches; it is difficult to find anything more
> hateful and inhumane in all of literature. If he wants to make his
> software free, fine. I've given away software I've written. No
> problem. But I think the concept of someone else telling me that my
> software should be free is going too far. I don't want to have anything
> to do with such hate-mongerers. If you think this is too extreme, do
> some research into the philosophy, software and otherwise, of the famous
> TLA-gurus. Be afraid. Be very afraid.
>
Sorry, what is TLA in this context and who are the "TLA-gurus"
--
Paul Sture
Robert Deininger
<jfmezei...@vaxination.ca> wrote:
>Ian Miller wrote:
>> ACMS is now supported by a HP team.
>
>
>What does "HP team" really mean ? Just people working out of home wherever
>they may be on this planet, or someone physically at ZKO ?
>
>Back in the days of Digital, was VMS engineering pretty much a physical
>team in a couple of locations ? (ZKP, Reading, Valbonnes and what other ?)
If you include VMS and layered products, Digital had people in multiple
locations in New Hampshire and Massachusetts, as well as contingents in
Colorado, several European countries, and India. And I'm sure I'm missing
some.
Michael Kraemer
>
> Simple. What differentiates VMS from other operating systems is the
> fact that it is well planned. Without meaning any harm to Linus, the
> fact that a student could write a unix kernel in his spare time says
> more about the lack of sophistication of the latter than the skills of
> the former.
No it says very little about anything.
The kernel is probably the easiest part of a Unix system,
in terms of man power. Probably of any OS.
(not that I would be able to write one, but a gifted
CS student should be)
It took about a decade and much more additional
man power until Linux was ready for serious use.
> I don't see how VMS could continue in an open-source world
> without a lot of coordination, and I don't see anyone with the knowledge
> and resources willing to do this for free. A lot of work is going into
> VMS now and this is funded by customers. Cut off the funding, cut out
> the development. Many of the customers would not be willing to pay for
> an open-source product.
Did you ask ?
> The big myth here is that somehow VMS is not everywhere because it is
> not free. Well, Windows is everywhere and it is not free. There are
> free operating systems which occupy a niche, and free ones with many
> users. And for hobbyists, VMS IS free. Are the masses signing up for
> DECUS memberships to get hobbyist licenses? No.
Well, compared to hobbyist use of other OSs,
it is still a PITA to deal with PAKs and annual renewal.
At least the latter is in argument that I heard more than once.
> And another danger of going the open-source route is the danger of
> approaching the RMS camp,
I see more danger in software patents, IP religion and DRM rip-off.
Much more.
> where not only should all software be free but
> people who write software for a living are brandmarked as morally
> inferior. Read some RMS speeches; it is difficult to find anything more
> hateful and inhumane in all of literature. If he wants to make his
> software free, fine.
RMS, that's Stallmann, I guess ?
Come on, you're taking that much too serious. Relax.
What I find annoying though, is that the Gnu stuff
has turned to be single platform only, basically.
In former times it was relatively easy to compile/port
on the platform of choice. Now it's practically Linux
only.
> Of course, HP owns VMS and HP makes money from VMS so even thinking
> about them open-sourcing it is idle speculation.
>
Wasn't it the common opinion in this group that HP doesn't want
to make money from VMS ?
Bill Gunshannon
Well, Digital used to have an office right here in Wilkes-Barre, PA.
But, according to the guy int he office next door, the CE's came in
one morning to find the locks on the door had been changed and they
were all bright enough to read the writting on the wall. Don't know
where they went, but I do know none of them still live around here.
(By the way, that being as it was still Digital then, it really was
quite soem time ago.)
bill
--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bi...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>
Bill Gunshannon
Michael Kraemer <M.Kr...@gsi.de> writes:
> Phillip Helbig---remove CLOTHES to reply schrieb:
>
>>
>> Simple. What differentiates VMS from other operating systems is the
>> fact that it is well planned. Without meaning any harm to Linus, the
>> fact that a student could write a unix kernel in his spare time says
>> more about the lack of sophistication of the latter than the skills of
>> the former.
>
> No it says very little about anything.
> The kernel is probably the easiest part of a Unix system,
Well, it's the easiest part to write badly.
> in terms of man power. Probably of any OS.
> (not that I would be able to write one, but a gifted
> CS student should be)
That can't be true or it would have left Linus out of the picture.
> It took about a decade and much more additional
> man power until Linux was ready for serious use.
Linux still isn't "ready for serious use". And it likely never will
be. But that won't stop people from doing it.
>
>> I don't see how VMS could continue in an open-source world
>> without a lot of coordination, and I don't see anyone with the knowledge
>> and resources willing to do this for free. A lot of work is going into
>> VMS now and this is funded by customers. Cut off the funding, cut out
>> the development. Many of the customers would not be willing to pay for
>> an open-source product.
>
> Did you ask ?
>
>> The big myth here is that somehow VMS is not everywhere because it is
>> not free. Well, Windows is everywhere and it is not free. There are
>> free operating systems which occupy a niche, and free ones with many
>> users. And for hobbyists, VMS IS free. Are the masses signing up for
>> DECUS memberships to get hobbyist licenses? No.
>
> Well, compared to hobbyist use of other OSs,
What other OSes? I am aware of no current OS that has a hobbyist
program that offers anything like VMS.
> it is still a PITA to deal with PAKs and annual renewal.
> At least the latter is in argument that I heard more than once.
I have never found that to be the case. The only problems I have
ever had were the timeing of the EDU license expiration (which really
has nothing to do with a Hobbyist License) and when I am carelees
enough to let mine expire without having asked for new ones.
>
>> And another danger of going the open-source route is the danger of
>> approaching the RMS camp,
>
> I see more danger in software patents, IP religion and DRM rip-off.
> Much more.
>
>> where not only should all software be free but
>> people who write software for a living are brandmarked as morally
>> inferior. Read some RMS speeches; it is difficult to find anything more
>> hateful and inhumane in all of literature. If he wants to make his
>> software free, fine.
>
> RMS, that's Stallmann, I guess ?
> Come on, you're taking that much too serious. Relax.
I am amazed to find someone who has read and understood RMS enough to
share my concern. The real scary part is that too many people already
take him "much too serious". If more people actually understood what
led to what is called the FSF today (and the true shortcomings of the
whole GNU concept from an RMS perspective) things would likely be much
different.
> What I find annoying though, is that the Gnu stuff
> has turned to be single platform only, basically.
> In former times it was relatively easy to compile/port
> on the platform of choice. Now it's practically Linux
> only.
Sound like anyone else we know? And people say proprietary OSes are
a bad thing!!
>
>> Of course, HP owns VMS and HP makes money from VMS so even thinking
>> about them open-sourcing it is idle speculation.
>>
>
> Wasn't it the common opinion in this group that HP doesn't want
> to make money from VMS ?
Doesn't mean there is a chance to get it open sourced. Just think
how lucky you are to just have a hobbyist program. Ask them to
allow hobbyist use of Ultrix-32 and see what happens.
Bob Koehler
>
> The kernel is probably the easiest part of a Unix system,
> in terms of man power. Probably of any OS.
The kernel in VMS is not so easy. It's quite complex compared to
the original time-sharing-only UNIX kernel. Part of that more
complex design is why it still has better realtime capabilities than
the UNIX kernels (like Tru64 and Solaris) that have had realtime
capabilities added (and are clearly more complex than the original).
Linux originally was also a pure timesharing kernel, and quite
simple. It's getting realtime pieces but I don't know if it can even
keep up with Solaris yet.
How easy a kernel is depends on how simple it's capabilties are.
Compare the above to Vxworks, which used a simple realtime-only
design and is just now adding the complexity needed for features
like memory protection that a timesharing or general purpose OS had
starting back in the '60s.
A general purpose OS, like VMS, which has both good timesharing and
good realtime capabilties, is not easy.
Bob Koehler
> What other OSes? I am aware of no current OS that has a hobbyist
> program that offers anything like VMS.
Thats what we need: a hobbyist program for Windows! Sure would
cut into Billy's pockets.
Seriously, friends tell me Solaris is pretty easy on their hobbyist
pocketbooks, too (free download). Tru64 was for a while at only $99,
but I think that's gone now.
Andrew
How about Solaris, its free to download for either SPARC or x86 and you
get the full release including compliers. You can also get the source
code via OpenSolaris.org. The only thnig Sun doesn't have is anything
called a hobbyist program however everything else is actually rather
better.
regards
Andrew Harrison
johnhre...@yahoo.com
Michael Kraemer wrote:
> Well, compared to hobbyist use of other OSs,
> it is still a PITA to deal with PAKs and annual renewal.
> At least the latter is in argument that I heard more than once.
Really? I have 14 systems that need to be renewed each year. I spent
about 20 minutes (or less) on the www.OpenVMSHobbyist.org site entering
in the serial numbers and then about another 20 minutes editing the
resulting e-mails into DCL command files (not that hard, they come
formatted as command files, I just pare them down a bit for what I
don't need) and applying them to the various machines. That's 40
minutes a year for 14 systems. Ouch. Please, spare me the pain ;-)
John H. Reinhardt
Bill Gunshannon
I don't think that counts as a "Hobbyist Program" as anyone can download
it. even businesses and they can use it for their business with Sun's
blessing. Small businesses have been able to do that since long before
this "open source" initiative.
Bill Gunshannon
"Andrew" <andrew_...@symantec.com> writes:
>
> Bill Gunshannon wrote:
>>
>> What other OSes? I am aware of no current OS that has a hobbyist
>> program that offers anything like VMS.
>
> How about Solaris, its free to download for either SPARC or x86 and you
> get the full release including compliers. You can also get the source
> code via OpenSolaris.org. The only thnig Sun doesn't have is anything
> called a hobbyist program however everything else is actually rather
> better.
My point, exactly. It isn't a "Hobbyist Program". Anyone can use it,
even businesses. I want to know what current, proprietary OSes offer
a Hobbyist Program and how it compares to the VMS Program. I am aware
of none, although I know of a number of platforms that have active
Hobbyist communities who would like to see programs.
hea...@aracnet.com
> Just to pick one aspect of such an idea, imagine what would happen to
> the VMS documentation.
<shudder!>
OK, that alone is reason enough to keep from going Open Source.
Zane
hea...@aracnet.com
> In article <emdd38$jsq$03$1...@news.t-online.com>,
> Michael Kraemer <M.Kr...@gsi.de> writes:
> > RMS, that's Stallmann, I guess ?
> > Come on, you're taking that much too serious. Relax.
> I am amazed to find someone who has read and understood RMS enough to
> share my concern. The real scary part is that too many people already
> take him "much too serious". If more people actually understood what
> led to what is called the FSF today (and the true shortcomings of the
> whole GNU concept from an RMS perspective) things would likely be much
> different.
I personally feel that Stallman does far more harm than good these days, and
has for years. I think part of the problem is his envy of the Linux kernel.
Linux is what he wanted Hurd to be.
I've run Linux since V0.12, and in the early days was quite active.
However, between Stallman and the "Rabid Fanboy" Linux crowd, I tend to stay
away from it these days, though still use it where it makes sense. Of
course I'm using Linux to type this, and just put two *BIG* Linux servers
into production yesterday, so I'm not staying that far away. At home mytwo
main systems run OpenVMS and Mac OS X.
> > What I find annoying though, is that the Gnu stuff
> > has turned to be single platform only, basically.
> > In former times it was relatively easy to compile/port
> > on the platform of choice. Now it's practically Linux
> > only.
> Sound like anyone else we know? And people say proprietary OSes are
> a bad thing!!
The problem is the people that write most of the code are familiar with two
OS's, Linux and Windows. Most have probably never touched another OS.
Personally, most things I write these days need to run on at least three
different operating systems.
Zane
ChrisQuayle
>
> I don't think that counts as a "Hobbyist Program" as anyone can download
> it. even businesses and they can use it for their business with Sun's
> blessing. Small businesses have been able to do that since long before
> this "open source" initiative.
>
> bill
>
>
Agreed. One of the many ways Sun's way of doing business did to improve
customer goodwill was that the os license came with the machine. For
example, if I bought a s/hand Sparc box, it was already licensed for the
version of Solaris that was current when it was originally sold. To
upgrade, just buy the new media kit. No lmf, hard work transfer
procedures and added lmf's and licenses for just about everything they
could think of, which was the Dec way of doing things. I don't think Sun
were or are particularly worried about os usage, as they see themselves
primarily as a hardware company, which was (is ?) where most of the
revenue was made. Come to think of it, Dec were orignally a hardware
company, so what went wrong ?.
It was petty, mean spririted stuff like the above that drove users away
from Dec in legions and even now, there appears to be little difference
after two changes of ownership. Technical excellence alone is not
enough, but must be combined with an ethical business policy for long
term success...
Chris
William Pechter
ChrisQuayle <nos...@devnul.co.uk> wrote:
>Bill Gunshannon wrote:
>
>>
>> I don't think that counts as a "Hobbyist Program" as anyone can download
>> it. even businesses and they can use it for their business with Sun's
>> blessing. Small businesses have been able to do that since long before
>> this "open source" initiative.
>>
>> bill
>>
>>
>
>Agreed. One of the many ways Sun's way of doing business did to improve
> customer goodwill was that the os license came with the machine. For
>example, if I bought a s/hand Sparc box, it was already licensed for the
>version of Solaris that was current when it was originally sold. To
>upgrade, just buy the new media kit. No lmf, hard work transfer
>procedures and added lmf's and licenses for just about everything they
>could think of, which was the Dec way of doing things. I don't think Sun
>were or are particularly worried about os usage, as they see themselves
>primarily as a hardware company, which was (is ?) where most of the
>revenue was made. Come to think of it, Dec were orignally a hardware
>company, so what went wrong ?.
>
Not true. To go to Solaris 2.3 from SunOS 4.1.3 and 4.1.4 I needed to
purchase a $300 upgrade license from Sun.
>It was petty, mean spririted stuff like the above that drove users away
>from Dec in legions and even now, there appears to be little difference
>after two changes of ownership. Technical excellence alone is not
>enough, but must be combined with an ethical business policy for long
>term success...
The problem is there were too many non-DECcies hired from places like
IBM with MBA degrees instead of Engineering degrees.
The problems began with the problem of not OEMing the VAX chips and
boards and killing their embedded systems customers... They went running
to Motorola and Zilog and Intel...
>
>Chris
>
>
Bill
--
--
"When I think back on all the crap I learned in Vax school
It's a wonder I fixed anything at all." (to the tune of Kodachrome)
pechter-at-ureach.com
Phillip Helbig---remove CLOTHES to reply
<paul.sture.nospam-1...@mac.sture.homeip.net>, Paul
Sture <paul.stu...@hispeed.ch> writes:
> Sorry, what is TLA in this context and who are the "TLA-gurus"
TLA := Three-Letter Acronym :-)
Phillip Helbig---remove CLOTHES to reply
> > Many of the customers would not be willing to pay for
> > an open-source product.
>
> Did you ask ?
I've met some.
> > The big myth here is that somehow VMS is not everywhere because it is
> > not free. Well, Windows is everywhere and it is not free. There are
> > free operating systems which occupy a niche, and free ones with many
> > users. And for hobbyists, VMS IS free. Are the masses signing up for
> > DECUS memberships to get hobbyist licenses? No.
>
> Well, compared to hobbyist use of other OSs,
> it is still a PITA to deal with PAKs and annual renewal.
> At least the latter is in argument that I heard more than once.
I spend about 5 minutes, TOTAL, renewing licenses each year.
> RMS, that's Stallmann, I guess ?
Yes.
> Come on, you're taking that much too serious. Relax.
I tended to take a "live and let live" approach until I by chance came
across transcriptions of some of his speeches. No matter why they do it
and what their motivation is, it is a harrowing experience to be branded
inferior, especially by someone who has never met me. Categorically
branding people as morally inferior, claiming that they have an agenda
which they don't etc just because they have a different opinion on
something is dangerous.
> Wasn't it the common opinion in this group that HP doesn't want
> to make money from VMS ?
What the common opinion in this group is and what HP actually does are
not always the same. :-|
Phillip Helbig---remove CLOTHES to reply
> >> The big myth here is that somehow VMS is not everywhere because it is
> >> not free. Well, Windows is everywhere and it is not free. There are
> >> free operating systems which occupy a niche, and free ones with many
> >> users. And for hobbyists, VMS IS free. Are the masses signing up for
> >> DECUS memberships to get hobbyist licenses? No.
> >
> > Well, compared to hobbyist use of other OSs,
>
> What other OSes? I am aware of no current OS that has a hobbyist
> program that offers anything like VMS.
Actually, I was thinking of you and Net BSD or whatever it's called. :-)
> I am amazed to find someone who has read and understood RMS enough to
> share my concern.
Likewise! The feeling is mutual!
Bill Gunshannon
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes:
> In article <4uvfmkF...@mid.individual.net>, bi...@cs.uofs.edu (Bill
> Gunshannon) writes:
>
>> >> The big myth here is that somehow VMS is not everywhere because it is
>> >> not free. Well, Windows is everywhere and it is not free. There are
>> >> free operating systems which occupy a niche, and free ones with many
>> >> users. And for hobbyists, VMS IS free. Are the masses signing up for
>> >> DECUS memberships to get hobbyist licenses? No.
>> >
>> > Well, compared to hobbyist use of other OSs,
>>
>> What other OSes? I am aware of no current OS that has a hobbyist
>> program that offers anything like VMS.
>
> Actually, I was thinking of you
Thanks. :-)
> and Net BSD or whatever it's called. :-)
But none of the BSD's have Hobbyist licenses. You can do anything
you want with then, including run your business.
>
>> I am amazed to find someone who has read and understood RMS enough to
>> share my concern.
>
> Likewise! The feeling is mutual!
>
>> The real scary part is that too many people already
>> take him "much too serious". If more people actually understood what
>> led to what is called the FSF today (and the true shortcomings of the
>> whole GNU concept from an RMS perspective) things would likely be much
>> different.
>
bill
Phillip Helbig---remove CLOTHES to reply
Gunshannon) writes:
> >> >> The big myth here is that somehow VMS is not everywhere because it is
> >> >> not free. Well, Windows is everywhere and it is not free. There are
> >> >> free operating systems which occupy a niche, and free ones with many
> >> >> users. And for hobbyists, VMS IS free. Are the masses signing up for
> >> >> DECUS memberships to get hobbyist licenses? No.
> >> >
> >> > Well, compared to hobbyist use of other OSs,
> >>
> >> What other OSes? I am aware of no current OS that has a hobbyist
> >> program that offers anything like VMS.
> >
> > Actually, I was thinking of you
>
> Thanks. :-)
>
> > and Net BSD or whatever it's called. :-)
>
> But none of the BSD's have Hobbyist licenses. You can do anything
> you want with then, including run your business.
He should have said "OS's which can be run at none or negligible costs by
hobbyists" as opposed to "hobbyist use of other OSs" to be completely
clear, though a hobbyist running an OS doesn't imply that that OS must
have a hobbyist license. It could, it could have a free license, it
could have no license, or it could have a commercial license (which
allows, but does not demand, commercial use).
Bill Gunshannon
But my point was to show that no one has offered a Hobbyist Program
that comes close to what we have with VMS. Most proprietary OSes do
not allow Hobbyist use under any circumstances. If you want to run
it, buy a commercial license (as has been pointed out in a recent
discussion in the PDP groups and for a while, cross-posted here.)
Others, even though no longer offered for sale commerically are still
specifically prohibited from hobbyist use (Primos). Legitimate Hobbyist
use is extremely rare and we have the best there is to offer!!
Paul Sture
bi...@cs.uofs.edu (Bill Gunshannon) wrote:
> In article <1166713928.1...@f1g2000cwa.googlegroups.com>,
> "Andrew" <andrew_...@symantec.com> writes:
> >
> > Bill Gunshannon wrote:
> >>
> >> What other OSes? I am aware of no current OS that has a hobbyist
> >> program that offers anything like VMS.
> >
> > How about Solaris, its free to download for either SPARC or x86 and you
> > get the full release including compliers. You can also get the source
> > code via OpenSolaris.org. The only thnig Sun doesn't have is anything
> > called a hobbyist program however everything else is actually rather
> > better.
>
> My point, exactly. It isn't a "Hobbyist Program". Anyone can use it,
> even businesses. I want to know what current, proprietary OSes offer
> a Hobbyist Program and how it compares to the VMS Program. I am aware
> of none, although I know of a number of platforms that have active
> Hobbyist communities who would like to see programs.
>
Well, not an OS, but you can download Oracle (Classic and Rdb) for
development purposes only.
--
Paul Sture
Phillip Helbig---remove CLOTHES to reply
Gunshannon) writes:
> But my point was to show that no one has offered a Hobbyist Program
> that comes close to what we have with VMS. Most proprietary OSes do
> not allow Hobbyist use under any circumstances. If you want to run
> it, buy a commercial license (as has been pointed out in a recent
> discussion in the PDP groups and for a while, cross-posted here.)
> Others, even though no longer offered for sale commerically are still
> specifically prohibited from hobbyist use (Primos). Legitimate Hobbyist
> use is extremely rare and we have the best there is to offer!!
One has to agree with that, of course!
Bob Koehler
> Come to think of it, Dec were orignally a hardware
> company, so what went wrong ?.
DEC separated the OS license from the hardware so that you didn't
have to buy both. So many people bought VAXen without a VMS
license that DEC looked into it and found they were all running BSD,
which lead them to market ULTRIX-32.
And back in PDP-11 days, you had your choice of several OS with
different capabilities. Its easier to sell you the one you want if
you list them as different parts, and it makes sense to charge more
for the more complex OS.
Even VAXen supported three DEC OS and a couple of non-DEC UNIX.
Sun didn't have such capabilities to offer, all they could do was
UNIX. No choices, so it might as well ship with the hardware.
Since they started out selling a BSD, there was no market users
to buy hardware from Sun and the OS from Berkley.
Main, Kerry
> -----Original Message-----
> From: hea...@aracnet.com [mailto:hea...@aracnet.com]
> Sent: December 21, 2006 2:46 PM
> To: Info...@Mvb.Saic.Com
> Subject: Re: Open sourced VMS as a business concept ?
>
> Bill Gunshannon <bi...@cs.uofs.edu> wrote:
> > In article <emdd38$jsq$03$1...@news.t-online.com>,
> > Michael Kraemer <M.Kr...@gsi.de> writes:
>
> > > RMS, that's Stallmann, I guess ?
>
[snip ...]
> > Sound like anyone else we know? And people say proprietary OSes are
> > a bad thing!!
>
> The problem is the people that write most of the code are
> familiar with two
> OS's, Linux and Windows. Most have probably never touched
> another OS.
> Personally, most things I write these days need to run on at
> least three
> different operating systems.
>
> Zane
>
Well, imho, the larger issue is the growing disconnect between those
writing applications and the various internal groups that need to
support them i.e. Operations, QA, Testing etc.
How many of those writing applications these days take into
consideration:
- monthly security patching which means re-testing important
applications with these security patches. How many App developers
actually even worry about OS or App security patches as that is the
"Operations" group worry?
- massive server consolidation initiatives which means longer distances
(latency, bandwidth issues etc) between end users and the servers new
target sites.
- much higher security and regulatory influences
- how to backup their applications in rapidly shrinking maint windows
e.g. lottieries and exchanges and other Internet global apps heading
towards 24x7x365 operations so App can not go down.
- How to ensure their Application does not call any HW specific calls
like time, physical devices etc that would make it extremely difficult
to virtualize their application?
- How to ensure their app is cluster aware so that multi-site solutions
can be easily implemented to handle their companies exponentially
increasing DT/DR/BC requirements.
Imho, the truly well designed app's of the future can no longer be
designed in a vacuum. They must take all of the above Operations and
Infrastructure issues into consideration. Unfortunately this will also
require much closer communications between the various groups and for
many companies that is a big challenge.
Btw, while Open Source and free OS's will continue to play a role in
some environments, I am also seeing a new trend developing in large
accounts where Senior Managers do not want their Technical staff playing
in the OS weeds every month, but rather spending more time talking to
the BU's and business groups to better understand their requirements and
communicate how IT can add value to their bottom lines.
Regards
Kerry Main
Senior Consultant
HP Services Canada
Voice: 613-592-4660
Fax: 613-591-4477
kerryDOTmainAThpDOTcom
(remove the DOT's and AT)
OpenVMS - the secure, multi-site OS that just works.
ChrisQuayle
>
> Not true. To go to Solaris 2.3 from SunOS 4.1.3 and 4.1.4 I needed to
> purchase a $300 upgrade license from Sun.
>
I paid around the same price for a media kit to go from 4.1.3 Sunos on a
3/60, to 4.1.4, which was the last of the pre sysv versions, but there
was nothing else to do in terms of compliance. No heavy duty implied
threat that paks came to represent. Paks were dec's way of saying that
we don't trust anyone, which really annoys. Completely the opposite
attitude to sun, who not only survived, but still look like they're
having fun doing it as well.
As you more or less say in your next paragraph, it's what happens when
you let you company get taken over by those with a passion for making
money, not a passion for engineering and doing stuff right.
Perhaps it's a sign of age, for those who remember the heady days of
moon landings etc, or is just another sign of the slow but seemingly
inevitable decline of western civilisation through the dilution and
corruption of basic values ?.
Sorry, didn't mean to get that deep, but have a good holiday in any case...
Chris
>
>
>>It was petty, mean spririted stuff like the above that drove users away
>
>>from Dec in legions and even now, there appears to be little difference
>
>>after two changes of ownership. Technical excellence alone is not
>>enough, but must be combined with an ethical business policy for long
>>term success...
>
>
> The problem is there were too many non-DECcies hired from places like
> IBM with MBA degrees instead of Engineering degrees.
>
> The problems began with the problem of not OEMing the VAX chips and
> boards and killing their embedded systems customers... They went running
> to Motorola and Zilog and Intel...
>
>
> Bill
ChrisQuayle
I can see the argument, but it's flawed. There's no reason why dec
couldn't have shipped any of the pdp os's with the box and taken a
liberal view about licensing. Imo, the real rot started when dec started
to see vax and vms as a high end system to rival mainframes and priced
everything accordingly. I well rememeber the uk dec direct catalog which
listed 3 x ra60's in a cabinet at an eye watering 66,000 uk pounds.
That's 22 k per drive !. They were *very* well engineered drives, but 22
k each ?...
Chris
roger
This might even be an extension of the hobbiest license where you could
get the latest copy of the source code and modify it. This could be
very different than the normal type of opened sourced projects like
Linux ---- Where the source code is available for people to look-at and
modify; but any changes would have to be submitted back to VMS
engineering where they would control and approve the changes. Even if
you had to give up the "rights" to the code; and agreed to not to
install alternate versions of VMS, etc. I know several people
extremely knowledgeable in VMS internals that just don't happen to work
for HP; and would submit bug fixes or changes now and then if they had
access and the ability to do this easily in a controlled way. We need
more young people learning and playing with VMS --- not less, and this
would be a way for people to learn VMS and improve upon it and learn.
Here is a few examples for just me in the last year. I needed some
global section information; but there is no system service to return
this information; so I wrote it. At first I wrote a routine to just
get the information I needed but then overtime I have improved it to
the point of a new system service routine. The new system service is
called sys$getgsi() -- to get global section information. There are
several parts of VMS that need this information internally (INSTALL
LIST/GLOBAL) for example or ANALY/SYS SHOW GSD but I needed it from
program. It's mostly finished and other people have even used it ---
and I would finish it and submit it if I knew it might become part of
VMS someday, but there is no good way to submit this and have it be
accepted or declined.
More recently, I need a lib$put_output_error(). The part of
sys$putmsg() that does output to either sys$output or sys$error if
different. Seems like this would be a good routine to have like
sys$put_output(). The code is buried in sys$putmsg() (somewhat ugly
the way it works with the SYS$PUTMSG logical name begin created), but
it would be nice to have a separate routine available for other people
to call and sys$putmsg could even call it.. But it's not there.
Another thing I could see happening overtime is that code would migrate
from BLISS or MACRO to C. Some people are going to yell at me for that
one, but people coming out of school know C and can work on C. All new
parts of VMS are in C, but there are many older parts still in macro or
bliss.
We have talked about this many times before, but I also think that a
port of VMS to AMD 64 or Intel IA32e is required -- this may even
require working with Intel or AMD to get all the mode bits and other
hardware requirements in a future version of the chip. I strongly
believe "that until VMS is ported to commodity PC hardware and until
VMS is available to be played with and learn from at the source code
level in some way by students -- it will continue to decline in
popularity and use..."
Michael Austin
and it is about time...
Most businesses also do not want opensource <OS, DB, anything> in their
production environments. When a mission/business critical system is down, who
are you going to call?
One thing that I continue to see is an attitude I have seen for a long time
coming from Windoze-weenies. "It is just a 1/2/4-CPU box". System designers
have truly lost the concepts that made VAX and Alpha systems far superior. The
concept that there is a lot more to the "system" than the processor. It's the
I/O stupid!! (and a OS that was truly designed - not thrown together as an
after-thought "oh, gee - we might want to have an OS for this new whiz-bang box"
On another similar topic, configuring a Sun active-active Oracle RAC cluster
takes 15-20 times longer on a two node cluster than configuring a 4- to 10-node
VMS Cluster running RDB.
--
Michael Austin.
Database Consultant
Alan Greig
> On another similar topic, configuring a Sun active-active Oracle RAC
> cluster takes 15-20 times longer on a two node cluster than configuring
> a 4- to 10-node VMS Cluster running RDB.
Which is a big plus if you're paid by the hour...
--
Alan Greig
William Pechter
Bob Koehler <koe...@eisner.aspm.encompasserve.org> wrote:
>In article <zpCih.20012$HV6....@newsfe1-gui.ntli.net>, ChrisQuayle
><nos...@devnul.co.uk> writes:
>
>> Come to think of it, Dec were orignally a hardware
>> company, so what went wrong ?.
>
> DEC separated the OS license from the hardware so that you didn't
> have to buy both. So many people bought VAXen without a VMS
> license that DEC looked into it and found they were all running BSD,
> which lead them to market ULTRIX-32.
>
> And back in PDP-11 days, you had your choice of several OS with
> different capabilities. Its easier to sell you the one you want if
> you list them as different parts, and it makes sense to charge more
> for the more complex OS.
>
> Even VAXen supported three DEC OS and a couple of non-DEC UNIX.
They pretty much had two I thought...
VAX/VMS
Ultrix-32
VAXELN was pretty much for target embedded apps... Don't think 11/7x0's
ran it.
>
> Sun didn't have such capabilities to offer, all they could do was
> UNIX. No choices, so it might as well ship with the hardware.
> Since they started out selling a BSD, there was no market users
> to buy hardware from Sun and the OS from Berkley.
>
Well, they did change that with the AT&T deal and the move to
Solaris2.x.
Also, NextStep was available for a while... and Solaris was (along with
WinNT) developed for PowerPC for a short time.
So the OS could be decoupled.
hea...@aracnet.com
> Also, NextStep was available for a while... and Solaris was (along with
> WinNT) developed for PowerPC for a short time.
> So the OS could be decoupled.
Wasn't that limited to the Sun Sparc 5, and not even all models of it? I
believe the number of S-Bus widgets supported was extremely small as well.
BTW, NeXTStep also supported at least one model of HP.
Zane
Arne Vajhøj
> Most businesses also do not want opensource <OS, DB, anything> in their
> production environments. When a mission/business critical system is
> down, who are you going to call?
????
I would have assumed that they would do exactly the same
as for a non open source software - they call whoever they
pay for support.
Arne
AEF
> In article <emdd38$jsq$03$1...@news.t-online.com>,
> Michael Kraemer <M.Kr...@gsi.de> writes:
> >
> >>
> >> Simple. What differentiates VMS from other operating systems is the
> >> fact that it is well planned. Without meaning any harm to Linus, the
> >> fact that a student could write a unix kernel in his spare time says
> >> more about the lack of sophistication of the latter than the skills of
> >> the former.
> >
> > No it says very little about anything.
> > The kernel is probably the easiest part of a Unix system,
>
> Well, it's the easiest part to write badly.
>
> > in terms of man power. Probably of any OS.
> > (not that I would be able to write one, but a gifted
> > CS student should be)
>
> That can't be true or it would have left Linus out of the picture.
>
> > It took about a decade and much more additional
> > man power until Linux was ready for serious use.
>
> Linux still isn't "ready for serious use". And it likely never will
> be. But that won't stop people from doing it.
Bill Gunshannon,
Can you please tell us why you think so poorly of Linux? I'm NOT saying
you're wrong as I am not in a position to judge. I'm just asking what
it is about Linux that makes you think so poorly of it.
Thanks.
[...]
>
> bill
>
> --
> Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
> bi...@cs.scranton.edu | and a sheep voting on what's for dinner.
> University of Scranton |
> Scranton, Pennsylvania | #include <std.disclaimer.h>
AEF
Bill Gunshannon
NIH Syndrome. Broken LPD implementation. Broken NFS implementation.
Extremely inefficient TCPIP implementation. Jus tto name a few.
How about this. Everytime I try to compile CMU's Festival Lite on a
Debian Linux box I have at home it crashes X. Want to take a guess at
how that happens? :-)
And we won't even get into the bogus license.
ChrisQuayle
>>
>>Can you please tell us why you think so poorly of Linux? I'm NOT saying
>>you're wrong as I am not in a position to judge. I'm just asking what
>>it is about Linux that makes you think so poorly of it.
>>
>>Thanks.
>>
>
>
> NIH Syndrome. Broken LPD implementation. Broken NFS implementation.
> Extremely inefficient TCPIP implementation. Jus tto name a few.
>
> How about this. Everytime I try to compile CMU's Festival Lite on a
> Debian Linux box I have at home it crashes X. Want to take a guess at
> how that happens? :-)
>
> And we won't even get into the bogus license.
>
> bill
>
My gripe is that it's becoming bloatware. If I install Debian Sarge on
an Alphastation 250/4/266, it's nearly an order of magnitude slower than
Tru64 on identical hardware. Slow to the point that it's almost
unusable. The same applies to Debian on Intel, compared to windows. Gut
feeling suggests that the culprits are the graphics subsystems, which
are starting to look far too complex and trivially decorative. If you
run Openwindows, rather than Gnome or similar, it's actually quite fast
and usable, which sort of confirms the suspicions. You need a powerull
machine to run Linux now - p3/500 is nowhere near good anough, which is
a long way from the original aims.
I don't contribute to the linux effort, nor have time to, so should
probably be gratefull, but would rather use Tru64 or Solaris, both of
which look a lot cleaner...
Chris
Bill Gunshannon
ChrisQuayle <nos...@devnul.co.uk> writes:
> Bill Gunshannon wrote:
>
>>>
>>>Can you please tell us why you think so poorly of Linux? I'm NOT saying
>>>you're wrong as I am not in a position to judge. I'm just asking what
>>>it is about Linux that makes you think so poorly of it.
>>>
>>>Thanks.
>>>
>>
>>
>> NIH Syndrome. Broken LPD implementation. Broken NFS implementation.
>> Extremely inefficient TCPIP implementation. Jus tto name a few.
>>
>> How about this. Everytime I try to compile CMU's Festival Lite on a
>> Debian Linux box I have at home it crashes X. Want to take a guess at
>> how that happens? :-)
>>
>> And we won't even get into the bogus license.
>>
>> bill
>>
>
> My gripe is that it's becoming bloatware.
Sadly, the bloat is creeping into the BSD's as well, thus my comment in
another thread about why it would be better to be able to run Ultrix-32
rather than NetBSD as HP recommends.
> If I install Debian Sarge on
> an Alphastation 250/4/266, it's nearly an order of magnitude slower than
> Tru64 on identical hardware. Slow to the point that it's almost
> unusable. The same applies to Debian on Intel, compared to windows. Gut
> feeling suggests that the culprits are the graphics subsystems, which
> are starting to look far too complex and trivially decorative. If you
> run Openwindows, rather than Gnome or similar, it's actually quite fast
> and usable, which sort of confirms the suspicions. You need a powerull
> machine to run Linux now - p3/500 is nowhere near good anough, which is
> a long way from the original aims.
Bloat totally aside, Linux is just too inefficient. One of the problems
with NIH Syndrome is they tend to ignore decades of research by the CSRG
and its follow-ons and just do things badly. Frequently making the same
mistakes others discovered and avoided long ago.
>
> I don't contribute to the linux effort, nor have time to, so should
> probably be gratefull, but would rather use Tru64 or Solaris, both of
> which look a lot cleaner...
Have you tried FreeBSD on your Alpha? I would be interested in hearing
what you thought and, especially, how it compares to Tru64. I alsready
know how it compares to Linux as I have run both on Alphas in the past.
David J Dachtera
...even if that's their own internal development group. I can't see undertaking
a move to local development without some local responsibility, open/closed
source not withstanding.
--
David J Dachtera
dba DJE Systems
http://www.djesys.com/
Unofficial OpenVMS Marketing Home Page
http://www.djesys.com/vms/market/
Unofficial Affordable OpenVMS Home Page:
http://www.djesys.com/vms/soho/
Unofficial OpenVMS-IA32 Home Page:
http://www.djesys.com/vms/ia32/
Unofficial OpenVMS Hobbyist Support Page:
http://www.djesys.com/vms/support/
hea...@aracnet.com
> Bill Gunshannon wrote:
> > NIH Syndrome. Broken LPD implementation. Broken NFS implementation.
> > Extremely inefficient TCPIP implementation. Jus tto name a few.
I'll bite, what is broken about LPD, and *which* LPD are you using. I'm
using LPRng to provide Unix printing to hundreds of printers and a bunch of
plotters for a wide variety of Unix clients.
> My gripe is that it's becoming bloatware. If I install Debian Sarge on
> an Alphastation 250/4/266, it's nearly an order of magnitude slower than
> Tru64 on identical hardware. Slow to the point that it's almost
> unusable. The same applies to Debian on Intel, compared to windows. Gut
> feeling suggests that the culprits are the graphics subsystems, which
> are starting to look far too complex and trivially decorative. If you
> run Openwindows, rather than Gnome or similar, it's actually quite fast
> and usable, which sort of confirms the suspicions. You need a powerull
> machine to run Linux now - p3/500 is nowhere near good anough, which is
> a long way from the original aims.
Linux has been suffering from *MAJOR* bloat for at least 10 years. The GUI
environment is defintely the biggest culprit. I'm running a copy of
Afterstep that is one of the pre-V1 beta's. I grabbed the source-code 10
years ago to run on AIX 3.2, and amazingly it still builds on current Linux
builds. Even Motif is lightweight compared to current Window managers!
See which will roll over and play dead first as you start putting a heavier
and heavier load on it, a currnet high-end PC running Linux, an older decent
PC running Solaris/x86, or an old low-end SGI system running a current
version of IRIX.
> I don't contribute to the linux effort, nor have time to, so should
> probably be gratefull, but would rather use Tru64 or Solaris, both of
> which look a lot cleaner...
Since IRIX is basically a dead platform now, Solaris is my Unix platform of
choice. Though these days I mainly use it to exercise scavanged HD's prior
to installing them on my VMS server.
I don't consider Mac OS X to be Unix, or even very efficient, however, I am
grateful for the Unix layer and X-Windows, as it keeps me from having to
turn on a Unix box most of the time.
Zane
Paul Sture
>
> Linux has been suffering from *MAJOR* bloat for at least 10 years. The GUI
> environment is defintely the biggest culprit. I'm running a copy of
> Afterstep that is one of the pre-V1 beta's. I grabbed the source-code 10
> years ago to run on AIX 3.2, and amazingly it still builds on current Linux
> builds. Even Motif is lightweight compared to current Window managers!
>
I'll agree on the GUI bloat in Linux. I used to have a dual boot system
with NT and Linux. The NT box was happy with 64 MB, better with 128 MB,
but to get a decent response from Linux I upgraded to 192 MB. This was
7-8 years ago when RAM still cost quite a lot.
>
> Since IRIX is basically a dead platform now, Solaris is my Unix platform of
> choice. Though these days I mainly use it to exercise scavanged HD's prior
> to installing them on my VMS server.
>
> I don't consider Mac OS X to be Unix, or even very efficient, however, I am
> grateful for the Unix layer and X-Windows, as it keeps me from having to
> turn on a Unix box most of the time.
>
Mac OS X also loves plenty of RAM. From what I others have said, if your
budget for a Mac system is limited, given the choice between 2 closely
systems, carefully consider the one where you can afford plenty of RAM.
MS X may not be a "proper" Unix, but I've personally found it an easier
introduction to that world than Linux.
Bill Gunshannon
Paul Sture <paul.stu...@hispeed.ch> writes:
> hea...@aracnet.com wrote:
>>
>> Linux has been suffering from *MAJOR* bloat for at least 10 years. The GUI
>> environment is defintely the biggest culprit. I'm running a copy of
>> Afterstep that is one of the pre-V1 beta's. I grabbed the source-code 10
>> years ago to run on AIX 3.2, and amazingly it still builds on current Linux
>> builds. Even Motif is lightweight compared to current Window managers!
>>
>
> I'll agree on the GUI bloat in Linux. I used to have a dual boot system
> with NT and Linux. The NT box was happy with 64 MB, better with 128 MB,
> but to get a decent response from Linux I upgraded to 192 MB. This was
> 7-8 years ago when RAM still cost quite a lot.
Well, let's be fair, the GUI is not really part of Linux. And it is the
same one that is used by the BSD's. Which brings us back to the question
"Why do the BSD's run so much better than Linux?" Guess the GUI isn't
the core of the problem after all. :-)
>
>>
>> Since IRIX is basically a dead platform now, Solaris is my Unix platform of
>> choice. Though these days I mainly use it to exercise scavanged HD's prior
>> to installing them on my VMS server.
>>
>> I don't consider Mac OS X to be Unix, or even very efficient, however, I am
>> grateful for the Unix layer and X-Windows, as it keeps me from having to
>> turn on a Unix box most of the time.
>>
>
> Mac OS X also loves plenty of RAM. From what I others have said, if your
> budget for a Mac system is limited, given the choice between 2 closely
> systems, carefully consider the one where you can afford plenty of RAM.
>
> MS X may not be a "proper" Unix, but I've personally found it an easier
> introduction to that world than Linux.
OS X is Unix. The only thing it offers of it's own is a Window Manager
that looks like theold Mac GUI. Everything else is and has been available
for other Unixes long before OS X decided to join the club.
Bob Koehler
> This could be
> very different than the normal type of opened sourced projects like
> Linux ---- Where the source code is available for people to look-at and
> modify; but any changes would have to be submitted back to VMS
> engineering where they would control and approve the changes.
This isn't all that different from what is really happening with
Linux. Linux is a kernel, it is owned by Linus and it contains what
he says it contains. Most of the stuff a user runs into is GNU
utilities that make a complete OS on top of that kernel. GNU's
license also calls for resubmission to GNU.
> Another thing I could see happening overtime is that code would migrate
> from BLISS or MACRO to C. Some people are going to yell at me for that
> one, but people coming out of school know C and can work on C. All new
> parts of VMS are in C, but there are many older parts still in macro or
> bliss.
Ouch! OK, I yelled at you. VMS Engineering has shown that it is
possible to write quality code in C, and if people submitted code
hopefully it would get the same quality reviews. But we've also
seen that rewriting code for no other reason than the language it's
in is a good place to introduce bugs. i.e. MAIL/OLD as a workaround
for MAIL rewritten in C.
Bob Koehler
> Bill Gunshannon wrote:
>> Linux still isn't "ready for serious use". And it likely never will
>> be. But that won't stop people from doing it.
>
> Bill Gunshannon,
>
> Can you please tell us why you think so poorly of Linux? I'm NOT saying
> you're wrong as I am not in a position to judge. I'm just asking what
> it is about Linux that makes you think so poorly of it.
Well I ONLY use it for serious purposes. In my experience its just
as ready as any UNIX.
ChrisQuayle
>
> Bloat totally aside, Linux is just too inefficient. One of the problems
> with NIH Syndrome is they tend to ignore decades of research by the CSRG
> and its follow-ons and just do things badly. Frequently making the same
> mistakes others discovered and avoided long ago.
Well, that's the human nature rebellious instinct that says "we can do
it better than you can" :-). Sometimes a great result, but as you say,
it helps to start with some idea of what works and what doesn't.
>
> Have you tried FreeBSD on your Alpha? I would be interested in hearing
> what you thought and, especially, how it compares to Tru64. I alsready
> know how it compares to Linux as I have run both on Alphas in the past.
Not as yet - hadn't really thought much about Freebsd yet, though the
Alpha support seemed little different to that of NetBsd/Alpha or even
Linux/Alpha last time I looked. There seems to be a lot of cross
fertilisation between the 3 now, which might explain your comments about
creeping bloat, but I guess it makes sense to share code to avoid
duplication of effort. Did play with some of the very early (v1.0'ish)
Netbsd vax releases a few years ago. What can be said ? - it works, but
the lack of framebuffer support for X limited its use at the time. As
for Linux, the poor perfomance of Linux/Alpha may be due to lack of work
to properly port the code to make full use of Alpha features, or even to
be 64 bit aware. One suspects that much of the Linux/Intel code has just
been recompiled for the Alpha port and who knows how good the compliers are.
The next step is to try Solaris on Intel and see how good that is at
dual P3/500 level. My guess is that it will rock solid and if so, there
will little reason to run Linux at all in the future.
So how does Freebsd compare with Linux, on Alpha, performance wise ?...
Chris
ChrisQuayle
> Since IRIX is basically a dead platform now, Solaris is my Unix platform of
> choice. Though these days I mainly use it to exercise scavanged HD's prior
> to installing them on my VMS server.
Ah yes, sun format, still about the best hd excerciser i've seen and
little changed since the days of Sun 3's. Recently picked up an A5100
and a load of 73 Gbyte fc drives to put in it. It was quite impressive
having 14 instances of format running in terminal windows at once and
all the drive leds flickering. Reminded me of *real* computers :-).
Long term, Solaris does indeed look like the platform of choice. One of
last of the commercial strength unices to survive and seems to get
better with every new release...
Chris
dav...@montagar.com
> Bill Gunshannon wrote:
>
> >
> > Bloat totally aside, Linux is just too inefficient. One of the problems
> > with NIH Syndrome is they tend to ignore decades of research by the CSRG
> > and its follow-ons and just do things badly. Frequently making the same
> > mistakes others discovered and avoided long ago.
>
> Well, that's the human nature rebellious instinct that says "we can do
> it better than you can" :-). Sometimes a great result, but as you say,
> it helps to start with some idea of what works and what doesn't.
Okay, maybe Linux isn't done as well as you think, Bill. Maybe
performance could be a little better, and with a little less bloat.
However, have you compared Linux with Windows on the same hardware?
Linux is faster. And performance for one MS app I do have? There is a
definite performance degradation for the same installation kit running
on Windows 2000 to Windows XP - with Windows XP running on faster
hardware with more memory! The Linux version beats them both on lesser
hardware. And now here comes Vista...
So I understand that you don't like Linux, but given the tasks I need
to perform and the tools available to perform them - OpenVMS runs my
server tasks (like Apache, MySQL, MX Mail, etc), and Linux provides the
desktop experience (Mail client, image manipulation, web browsing, code
development, documentation, etc). And my code ports to Windows and
OpenVMS, too - because I use standard and available libraries to insure
that.
So maybe Linux is a pair of pliers when you would prefer a finely
crafted Snapon box wrench. But I have Linux, and it does work. There
isn't the right size of Snapon box wrench available, and the Windows
version is, well, let's just say it's not a Craftsman, either.
Bill Gunshannon
ChrisQuayle <nos...@devnul.co.uk> writes:
> Bill Gunshannon wrote:
>
>>
>> Bloat totally aside, Linux is just too inefficient. One of the problems
>> with NIH Syndrome is they tend to ignore decades of research by the CSRG
>> and its follow-ons and just do things badly. Frequently making the same
>> mistakes others discovered and avoided long ago.
>
> Well, that's the human nature rebellious instinct that says "we can do
> it better than you can" :-). Sometimes a great result, but as you say,
> it helps to start with some idea of what works and what doesn't.
I guess I don't see how making the same mistakes over again is going to
result in a better product. And with Linux as the example, it obviously
doesn't.
>
>>
>> Have you tried FreeBSD on your Alpha? I would be interested in hearing
>> what you thought and, especially, how it compares to Tru64. I alsready
>> know how it compares to Linux as I have run both on Alphas in the past.
>
> Not as yet - hadn't really thought much about Freebsd yet, though the
> Alpha support seemed little different to that of NetBsd/Alpha or even
> Linux/Alpha last time I looked. There seems to be a lot of cross
> fertilisation between the 3 now, which might explain your comments about
> creeping bloat, but I guess it makes sense to share code to avoid
> duplication of effort.
There is some, but the big difference is that while the {Net|Open}BSD
crew try to make an OS that runs on everything the FreeBSD crew decided
early on to concentrate on the most likely targets and have doen some
optimization in that direction.
> Did play with some of the very early (v1.0'ish)
> Netbsd vax releases a few years ago. What can be said ? - it works,
"Works" is a matter of opinion. It won't use modern hardware like DSSI or
CI so the choice of disks is pretty slim (SCSI being the best bet if you
can find one and can afford it). It needs lots of memory. More than you
will find on most smaller VAXen, like 3100's or the MicroVAX. Last time I
tried a kernel rebuild on one of my old MicroVAX-II's here I finally gave
up and killed it after more than 48 hours. I love my VAXen, but BSD has
grown beyond usability on one.
> but
> the lack of framebuffer support for X limited its use at the time.
That too. :-) Even the Linux guys gave up on the VAX long before they
had anything viable.
> As
> for Linux, the poor perfomance of Linux/Alpha may be due to lack of work
> to properly port the code to make full use of Alpha features, or even to
> be 64 bit aware. One suspects that much of the Linux/Intel code has just
> been recompiled for the Alpha port and who knows how good the compliers are.
Well, try runing some benchmarks on Linux and FreeBSD on the same 32bit
hardware. It is not specificly 64bit code that is inefficient. But then,
if they started with garbage, moving to 64bits wasn't likely to magically
make things better.
>
> The next step is to try Solaris on Intel and see how good that is at
> dual P3/500 level. My guess is that it will rock solid and if so, there
> will little reason to run Linux at all in the future.
Yeah, it is. But it is rapidly morphing in such a way that I suspect
before very long you will need special skills to maintain Solaris
systems and software incompatabilities will begin to creep in again.
I honestly think they should have stayed with SunOS and just kept
modernizing it but the marketing hyoe said that SYSV was the wave of
future. Kind of like Itanium.
>
> So how does Freebsd compare with Linux, on Alpha, performance wise ?...
I haven't run Linux on an Alpha in a really long time. FreeBSD performed
much better the las ttime I did and I have seen nothing in the Linux camp
to make me think this has changed in any way other than the possibility
that FreeBSD has gotten even better.
I have to keep up with Linux because DA has settled on it after abandoning
Solaris. Personally, I can't for the life of me imagine why they chose
Linux over FreeBSD except possibly because of all the hype surounding
Linux and the total silence surrounding any of the BSD's.
Bill Gunshannon
dav...@montagar.com writes:
> ChrisQuayle wrote:
>> Bill Gunshannon wrote:
>>
>> >
>> > Bloat totally aside, Linux is just too inefficient. One of the problems
>> > with NIH Syndrome is they tend to ignore decades of research by the CSRG
>> > and its follow-ons and just do things badly. Frequently making the same
>> > mistakes others discovered and avoided long ago.
>>
>> Well, that's the human nature rebellious instinct that says "we can do
>> it better than you can" :-). Sometimes a great result, but as you say,
>> it helps to start with some idea of what works and what doesn't.
>
> Okay, maybe Linux isn't done as well as you think, Bill. Maybe
> performance could be a little better, and with a little less bloat.
> However, have you compared Linux with Windows on the same hardware?
Who cares about Windows?
> Linux is faster.
And FreeBSD is faster still.
> And performance for one MS app I do have? There is a
> definite performance degradation for the same installation kit running
> on Windows 2000 to Windows XP - with Windows XP running on faster
> hardware with more memory! The Linux version beats them both on lesser
> hardware. And now here comes Vista...
Like I said, who cares about Windows? Let's stick with apples and apples.
I was comparing Unix and a really bad Unix clone. Oh yeah, and we were
talking about running on an Alpha. You got Xp for the Alpha? Hoe about
Vista? I didn't think so.
>
> So I understand that you don't like Linux, but given the tasks I need
> to perform and the tools available to perform them - OpenVMS runs my
> server tasks (like Apache, MySQL, MX Mail, etc), and Linux provides the
> desktop experience (Mail client, image manipulation, web browsing, code
> development, documentation, etc).
And all the desktp apps that run on Linux run on FreeBSD, better. At
least on BSD a user (not root) can't bring down teh whole GUI by running
the C compiler. (Yes, I have done it a number of time recently and finaly
had to give up trying to build the application on Linux entirely.)
> And my code ports to Windows and
> OpenVMS, too - because I use standard and available libraries to insure
> that.
Which has waht to do with the comparison between Linux and FreeBSD on
an Alpha? If, as you say, your code uses "standard and available libraries"
it would obviously run on FreeBSD as well. Most likely, depending on what
they actually do, better.
>
> So maybe Linux is a pair of pliers when you would prefer a finely
> crafted Snapon box wrench. But I have Linux, and it does work. There
> isn't the right size of Snapon box wrench available, and the Windows
> version is, well, let's just say it's not a Craftsman, either.
I really don't understand what Windows had to do witht he discussion
that was taking place.
A better comparison would be FreeBSD = wrench vs. Linux = a boxing glove.
Andrew
Bob Koehler wrote:
> In article <emdd38$jsq$03$1...@news.t-online.com>, Michael Kraemer <M.Kr...@gsi.de> writes:
> > Phillip Helbig---remove CLOTHES to reply schrieb:
> >
> > The kernel is probably the easiest part of a Unix system,
>
> the original time-sharing-only UNIX kernel. Part of that more
> complex design is why it still has better realtime capabilities than
> the UNIX kernels (like Tru64 and Solaris) that have had realtime
> capabilities added (and are clearly more complex than the original).
>
I don't know where you got your information from but you are incorrect
at least with respect to Solaris.
The original Solaris 2.0 kernel was designed to support real-time with
capabilities such as full thread pre-emption. True some of the user
land interfaces such as POSIX real-time support took longer to arrive
but Solaris itself was designed to support real-time.
http://citeseer.ist.psu.edu/update/1716
Regards
Andrew Harrison
> Linux originally was also a pure timesharing kernel, and quite
> simple. It's getting realtime pieces but I don't know if it can even
> keep up with Solaris yet.
>
> How easy a kernel is depends on how simple it's capabilties are.
> Compare the above to Vxworks, which used a simple realtime-only
> design and is just now adding the complexity needed for features
> like memory protection that a timesharing or general purpose OS had
> starting back in the '60s.
>
> A general purpose OS, like VMS, which has both good timesharing and
> good realtime capabilties, is not easy.