Re: FW: WTT: 1.5G of PC2700 for 1G of PC100

[Jochen Kunz]

On Mon, 5 May 2008 17:30:39 -0400
Lionel Peterson <lionel4287@...> wrote:

>   Too bad that
> > NetBSD doesn't run on the WRT54GLs.  If it did, I could move one to
> > a rack case, and run long antenna leads to place one antenna on each
> > floor (maybe that is a bad idea, I obviously don't know much about  
> > RF).
>
> Bad idea IMHO.
>
> The RF leakage would be excessive unless you secured high-quality,
> low- loss wire for the runs...

I would not be to much worried about RF leakage. The problem is that
even high quality "low loss" cable has a fairly high loss in the 2.5
GHz range. Not to talk about the cost of that cable and proper
connectors... In other words: It is cheaper and more effective to have
multiple accesspoints and place them at the point where they are needed.
--


tsch__,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Peter Corlett]

On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
> Shannon Hendrix wrote:
[...]
>> You can get a comp-sci degree without ever leaving Netbeans...
> You can get a comp-sci degree without ever having touched a computer.

That's because Computer Science is not a vocational qualification.

A decent CS degree will have lots of hands-on work from the gate level up,
of course.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Peter Corlett]

On Mon, May 05, 2008 at 06:51:59AM +0300, Geoffrey S. Mendelson wrote:
> On Sun, May 04, 2008 at 08:14:22PM -0600, Robert Darlington wrote:
[...]
>> Assemblers assemble assembly. Sorry, couldn't resist.
> At least you did not say that they COMPILE it. I've heard that too many
> times. :-)

They get quite close sometimes in that they can expand macros, evaluate
expressions and do peephole optimisations.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Phil Stracchino-3]

Peter Corlett wrote:

> On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
>> Shannon Hendrix wrote:
> [...]
>>> You can get a comp-sci degree without ever leaving Netbeans...
>> You can get a comp-sci degree without ever having touched a computer.
>
> That's because Computer Science is not a vocational qualification.
>
> A decent CS degree will have lots of hands-on work from the gate level up,
> of course.

That.  In my CS course at Eastern Washington, we did all the theory,
some familiarization with different OSen and some coding in languages,
but we also built (well, simulated) chips from the gate level and built
prototype hardware.  (What my CS class designed and built is used today
by NASA.)


--
   Phil Stracchino, CDK#2     DoD#299792458     ICBM: 43.5607, -71.355
   alaric@...   alaric@...   phil@...
          Renaissance Man, Unix ronin, Perl hacker, Free Stater
                  It's not the years, it's the mileage.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Sridhar Ayengar]

Jochen Kunz wrote:
> On Mon, 5 May 2008 15:46:47 -0400
> Joshua Boyd <jdboyd@...> wrote:
>
>> I also like NetBSD, but I don't believe that the ZFS port is usable
>> yet.  
> No ZFS in NetBSD yet. And I will not hold my breath waiting for it.

Isn't the FreeBSD ZFS code approaching production-quality?

> On the other side: The soon to be branched 5.0 will have good, old, rock
> solid FFS with shining new jornaling support...

Just for metadata or for data too?

Peace...  Sridhar
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Sridhar Ayengar]

Peter Corlett wrote:
> On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
>> Shannon Hendrix wrote:
> [...]
>>> You can get a comp-sci degree without ever leaving Netbeans...
>> You can get a comp-sci degree without ever having touched a computer.
>
> That's because Computer Science is not a vocational qualification.

That doesn't stop headhunters, hiring managers and HR departments from
treating it as one.

Peace...  Sridhar
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Robert Darlington]

Hah, at my old job you had to have a degree in a science as a major
qualifying factor in becoming a technical staff member.  So, to get
the best pay as a systems administrator, you needed to have a degree
in any science, be it computer science, physics, chemistry, or just
about anything else.  An MIS degree would get you nothing, even though
we managed information systems and didn't do computer science.  I
can't remember the last day as a sysadmin that I needed to
mathematically describe the fastest way to sort a filing cabinet if it
had 58 folders in it.  I personally feel a CS grad would make a great
secretary or technical writer for procedures based texts, but not a
sysadmin.

On a side note, are there any universities in the States that still
teach C?  Mine (UNM) has C++ as an elective and uses Java as the main
focus for teaching CS, but no plain ol' C.  The only other languages
they teach are object oriented PERL and VB, both as electives.  This
just goes to show that CS has very little to do with being a
programmer anymore.

-Bob

On Tue, May 6, 2008 at 9:18 AM, Sridhar Ayengar <ploopster@...> wrote:

> Peter Corlett wrote:
>
> > On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
> >
> > > Shannon Hendrix wrote:
> > >
> > [...]
> >
> > >
> > > > You can get a comp-sci degree without ever leaving Netbeans...
> > > >
> > > You can get a comp-sci degree without ever having touched a computer.
> > >
> >
> > That's because Computer Science is not a vocational qualification.
> >
>
>  That doesn't stop headhunters, hiring managers and HR departments from
> treating it as one.
>
>  Peace...  Sridhar
>
>
>  _______________________________________________
>  rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Tue, May 06, 2008 at 04:26:09PM +0200, Jochen Kunz wrote:
> On Mon, 5 May 2008 15:46:47 -0400
> Joshua Boyd <jdboyd@...> wrote:
>
> > I also like NetBSD, but I don't believe that the ZFS port is usable
> > yet.  
> No ZFS in NetBSD yet. And I will not hold my breath waiting for it.

There are licensing problems, aren't there?  Or is it just dead in the
water?

I knew it wasn't in NetBSD, but I thought it was most of the way there
from the GSoC project.
 
> On the other side: The soon to be branched 5.0 will have good, old, rock
> solid FFS with shining new jornaling support...

That might help.  I'll be watching what people say, but I doubt that it
will cause me to consider switching from Solaris for a file server.
Especially as long as there continue to be SMP issues on SPARC64
(although I keep seeing encouraging progress).
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson]

>From: Jochen Kunz <jkunz@...>
>Date: 2008/05/06 Tue AM 09:32:52 CDT
>To: rescue@...
>Subject: Re: [rescue] FW: WTT: 1.5G of PC2700 for 1G of PC100

>On Mon, 5 May 2008 17:30:39 -0400
>Lionel Peterson <lionel4287@...> wrote:
>
>>   Too bad that
>> > NetBSD doesn't run on the WRT54GLs.  If it did, I could move one to
>> > a rack case, and run long antenna leads to place one antenna on each
>> > floor (maybe that is a bad idea, I obviously don't know much about  
>> > RF).
>>
>> Bad idea IMHO.
>>
>> The RF leakage would be excessive unless you secured high-quality,
>> low- loss wire for the runs...
>I would not be to much worried about RF leakage. The problem is that
>even high quality "low loss" cable has a fairly high loss in the 2.5
>GHz range. Not to talk about the cost of that cable and proper
>connectors... In other words: It is cheaper and more effective to have
>multiple accesspoints and place them at the point where they are needed.

Exactly - my concern wasn't a health issue, it was about getting a resonable signal to the antenna over an extended run of coax... Most access points emit less than 100 mW, some much less, and pose minimal (if any) health risks.

Lionel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson]

>From: Peter Corlett <abuse@...>
>Date: 2008/05/06 Tue AM 09:45:52 CDT
>To: The Rescue List <rescue@...>
>Subject: Re: [rescue] Small servers (was Re: WTT: 1.5G of PC2700 for 1G of PC100)

>On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
>> Shannon Hendrix wrote:
>[...]
>>> You can get a comp-sci degree without ever leaving Netbeans...
>> You can get a comp-sci degree without ever having touched a computer.
>
>That's because Computer Science is not a vocational qualification.
>
>A decent CS degree will have lots of hands-on work from the gate level up,
>of course.

I attended (but did not graduate from) a small liberal arts college in central PA about 25 years ago, and the BIG DEBATE among the COmp Sci faculty was if they were going to teach COBOL or not to kids majoring in Comp Sci. The debate focused on the *perception* of reducing the degree to a certificate from a vocational school if they did teach COBOL.

The counter argument was that without some grounding in commercial computer technologies, it was exceedingly difficult to get a job with your four year degree in Comp Sci. Back then, C, C++, and 68000 assembler were the main languages taught, with some BASIC for non-Comp Sci majors in the Intro classes. Java hadn't been invented yet, and people were over Pascal already in most situations (as I recall).

I tried to convince the department heads that it would make sense to offer COBOL, FORTRAN, and some mainframe topics as "seminars" outside the normal course work for those students interested (not graded, not required for graduation).

Lionel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jochen Kunz]

On Tue, 06 May 2008 11:17:15 -0400
Sridhar Ayengar <ploopster@...> wrote:

> > No ZFS in NetBSD yet. And I will not hold my breath waiting for it.
> Isn't the FreeBSD ZFS code approaching production-quality?
FreeBSD != NetBSD. IIRC the last year GSoC ZFS for NetBSD project got as far as a mostly working ZVol layer and will be continued this year.

> > On the other side: The soon to be branched 5.0 will have good, old,
> > rock solid FFS with shining new jornaling support...
> Just for metadata or for data too?
Metadata only. See:
http://mail-index.netbsd.org/tech-kern/2008/03/02/msg000504.html

If you need data-journaling LFS would be the better route.
--


tsch__,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jochen Kunz]

On Tue, 6 May 2008 13:33:55 -0400
Joshua Boyd <jdboyd@...> wrote:

> > No ZFS in NetBSD yet. And I will not hold my breath waiting for it.
> There are licensing problems, aren't there?  Or is it just dead in the
> water?
Worst case would be that ZFS is only available as LKM to keep CDDL code
out of the kernel source tree.

> I knew it wasn't in NetBSD, but I thought it was most of the way there
> from the GSoC project.
AFAIK FreeBSD got quite a bit further with the ZFS integration.

> > On the other side: The soon to be branched 5.0 will have good, old,
> > rock solid FFS with shining new jornaling support...
> That might help.  I'll be watching what people say, but I doubt that
> it will cause me to consider switching from Solaris for a file server.
> Especially as long as there continue to be SMP issues on SPARC64
> (although I keep seeing encouraging progress).
Sure. If you have sparc64 hardware, you are used to Solaris and you are
already running Solaris, why switch to an other OS if there is no
immediate need? Never change a runing system! :-)
Especially as sparc64 hardware is designed to run Solaris and Solaris
is designed to run on sparc64 hardware.
--


tsch__,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Sridhar Ayengar]

Jochen Kunz wrote:
> On Tue, 06 May 2008 11:17:15 -0400
> Sridhar Ayengar <ploopster@...> wrote:
>
>>> No ZFS in NetBSD yet. And I will not hold my breath waiting for it.
>> Isn't the FreeBSD ZFS code approaching production-quality?
> FreeBSD != NetBSD. IIRC the last year GSoC ZFS for NetBSD project got as far as a mostly working ZVol layer and will be continued this year.
   ^^^^^^^^^^^^^^^^^
No, but architecturally, it's closer than Linux or Solaris are.  It
wouldn't surprise me if some of the work could be leveraged.

>>> On the other side: The soon to be branched 5.0 will have good, old,
>>> rock solid FFS with shining new jornaling support...
>> Just for metadata or for data too?
> Metadata only. See:
> http://mail-index.netbsd.org/tech-kern/2008/03/02/msg000504.html
>
> If you need data-journaling LFS would be the better route.

But then LFS doesn't have metadata journaling, does it?  And is LFS (as
of 4.0) now ready for prime-time?

Peace...  Sridhar
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jochen Kunz]

On Wed, 07 May 2008 15:16:26 -0400
Sridhar Ayengar <ploopster@...> wrote:

> But then LFS doesn't have metadata journaling, does it?
LFS doesn't have journaling at all. It can't have journaling because it
is not a journaling FS, it is log structured. (Like NetApps WAFL or
ZFS. Yes, AFAIK ZFS is log structured, not journaled.) Due to the
inherent way LFS works you get the same effect as on a data +
metadata journaling FS: Either the last write succeded and you get the latest data + metadata. Or the last write didn't complete and you get the previous data + metadata.

> And is LFS (as of 4.0) now ready for prime-time?
Well. To be onest. If I had important data on LFS I would ensure to
have a good backup... All I can say is: I tried it at some occasions on
sparc64 and it worked well for me.
--


tsch__,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 5, 2008, at 12:41 , William Enestvedt wrote:

> So, getting back to the original question....Shannon, you pick  
> anything
> yet to use as a small server? I'm interested in this question, too.

Not yet, but I'm looking at the Alix stuff and other things.

I have even thought of another PC server with a smaller case than the  
"big" Dell T105, but was hoping to find something *REALLY* small if  
possible, and use a notebook drive for storage.

The problem with some of the small stuff is the CPUs are a little  
wimpy, and while DNS does not care, squid does, and there is even a  
possibility I'll put email on it.

It would be nice if the Solaris machine (the T105) was more free for  
being shut down for maintenance and software changes.

Plus, admittedly, I think tiny servers are really cool.

I've always wondered why someone doesn't make a case that fits around  
hardware, insteads of the mostly empty boxes we typically get.

There is no reason that something the size of a Shuttle could not have  
a full RAID bay in it, for example.

I wish that whole job thing was not getting in the way of my research  
on this project... :)

When I come up with something I'll take pictures and documented it and  
all that good stuff.  Just taking my time since I can do without this  
one for now.



--
Shannon Hendrix
shannon@...
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 5, 2008, at 10:14 , der Mouse wrote:

>>> Which is as it should be; computer science is a theoretical  
>>> discipline
>> No, it is not.  It is an applied science just like medicine or
>> engineering.
>
> That's a substantial part of the problem, then; different people
> understanding the same words to mean different things.
>
> What do _you_ call the theoretical discipline, then?

Mathematics?

The problem with "computer science" is that it is very difficult to  
learn anything useful without using a real machine, just as it is  
quite difficult to learn anything useful in medicine if you never  
touch a test tube, a patient, or chemicals.

Sure, you could "theorize" about gall bladders, but that doesn't tend  
to do the world very much good.

In fact, it is quite rare for even theoreticians to be purely  
theoreticians if they want their work to be at all useful.

> The one to which
> belong things like the proof that general sorting cannot happen faster
> than n log(n), the Halting Problem, et al?

Seems to me that "Computer Science" is a combination of theory and  
engineering, just like most useful learning.

Why do you feel the need to draw a line?

It's almost useless to do that with pure math and physics, and  
definitely useless for computer science.

At some point, you have to demonstrate that you are right about your  
O(n) predictions (which are often quite wrong on real hardware), or to  
demonstrate exactly why some problems are np complete, etc.

It's all well and good the theorists have speculated that there is a  
world out there but they'd all starve to death of the scientists and  
engineers had not lived in that world and done things with it.




--
Shannon Hendrix
shannon@...
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 5, 2008, at 16:00 , Robert Darlington wrote:

> And absolutely none of that applies to high performance computing.
> Try running object oriented code where time counts!  I guess you could
> always throw more computing power at a problem, but the problems I've
> worked on take months to solve with computers that cost half a billion
> dollars.  Going OO would increase the time to solution many times
> over.  Fortran 77 for life, pointers be damned.

One of the best performance charts I have ever seen is in "Programming  
Pearls".

They show a TRS-80 and a DEC Alpha doing the same job, but the TRS-80  
code is better written.

The DEC is hundreds of times faster... until the problem size  
increases to a point at which the DEC would take 400 years to do what  
the TRS-80 does in fifteen minutes.

Excellent example of how little raw power means if you use it wrong,  
and exactly why throwing horsepower or misguided ideology at a problem  
is stupid.

Programmers do *NOT* always cost more than computers.

--
"Where some they sell their dreams for small desires."
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 5, 2008, at 16:15 , Ethan O'Toole wrote:

>> That's just a video->serial conversion.
>> I mean a BIOS designed to talk to serial devices, not a video  
>> converter.
>
> The console redirection is actually a function where it sees the  
> characters being written to the video BIOS and pushes them out of a  
> serial port.

On some systems, sure, but not all of them.

> It is not a screen scrape.

It is if the BIOS does direct video writes, and several of them do.

> Anything that writes directly to the video card to speed up display  
> will bypass PC serial consoles.

I have several machines where their console redirect is implemented on  
the video card.

The bottom line is we need a BIOS that puts a command line up on a  
real serial console, no video or BIOS translation, but real console  
support.

> I don't mind the PC stuff really. I mean, once a server is setup and  
> deployed you don't really have to deal with it that much. Even on  
> Sun and SGI boxes you don't really mess with the boot prom stuff  
> once it's up and running.

The issue comes at 0400 when you have to drive out to a server, or  
even fly out to it, all because they could not spend $1.50 in quantity  
to have a real console.

Even at home, it's trivial to use a serial console box to work with  
several boxes, but a bulky and ugly mess to do it with video.


--
"Where some they sell their dreams for small desires."
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 6, 2008, at 11:18 , Sridhar Ayengar wrote:

> Peter Corlett wrote:
>> On Sun, May 04, 2008 at 08:44:56PM -0400, Sridhar Ayengar wrote:
>>> Shannon Hendrix wrote:
>> [...]
>>>> You can get a comp-sci degree without ever leaving Netbeans...
>>> You can get a comp-sci degree without ever having touched a  
>>> computer.
>> That's because Computer Science is not a vocational qualification.
>
> That doesn't stop headhunters, hiring managers and HR departments  
> from treating it as one.

They don't want a person trained to learn, they want a person trained  
to not learn.

Personally, I think "Computer Science" means a combination of theory  
and practice, and a  heavy focus on learning how to learn and  
foundations which you eventually come back and expand on and/or move  
into vocational training for specific needs.

The degree itself certainly is not and should not be vocational.  
Vocational training is narrowly focused and limited, and usually means  
you have to retrain to learn something else.

By contrast a Computer Science degree should teach you how to learn  
almost any related technology or create your own, or know what  
training you need to get there.  You are supposed to be prepared for a  
life of learning, not just "getting a job".

The problem is the modern workplace just does not want people to  
learn, they want to plug square people into square holes and be done  
with it.

--
"Where some they sell their dreams for small desires."
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Shannon Hendrix]

On May 6, 2008, at 12:06 , Robert Darlington wrote:

> I personally feel a CS grad would make a great
> secretary or technical writer for procedures based texts, but not a
> sysadmin.

Well, I'm comp-sci and do a better job with admin than most of the  
admins I have met.  I frequently have to do my own admin work if I  
want anything to work right.

A good computer science degree is a superset of what you need for  
being a good system administrator.  It's mostly a matter of do you  
want to be an admin, or would you rather stick to more traditional  
comp-sci jobs.

It is more efficient in many cases to separate those jobs, but only if  
each side is competent and understands the needs of the other.

Most sysadmins I've worked with who just had vocational training are  
terrible.  They might really know the material they were taught, but  
have a really hard time when you get away from that.  Worse, it is  
impossible to explain to them why some of their decisions are a  
disaster for the software and running it efficiently.

Certainly an admin can self-