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

[wa2egp]

> 1. Always initialize variables.
>
> 2. Validate the parameters to a system call before you call it.
>
> 3. Validate the results of a system call (as in don't use a NULL pointer).
>
> 4. Always make sure the data fits in the space allocated for it.
>
> 5. Always make sure an I/O operation succeeds.
>
> 6. Put lots of debugging statments in your code.
>    
> 7. Comments, comments, comments.
>  
> 8. Write as simple code as possible. If it is difficult for someone to
>    understand, it will be impossible to maintain. (KISS).
>
> Geoff.

Also remember constants aren't and variable won't.  :)

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

[Jonathan C. Patschke]

On Mon, 5 May 2008, 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.

I have to agree with you.  One of the industry's biggest problem is that
students and employees both think that the computer science discipline
is something it isn't.  People go through a CS program expecting to know
how to implement systems instead of how to design the abstract parts of
them.

If folks want an electrical/computer engineer, they should hire one.

--
Jonathan Patschke | "There is more to life than increasing its speed."
Elgin, TX         |                                   --Mahatma Gandhi
USA               |
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[William Enestvedt]

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.

    There's a big ol' tower PC (with WinXP) in my basement I would like
to throw out -- well, I just want to be rid of the bag power draw,
really. The slower (c. 800 MHz?) CPU doesn't bug me much, and I don't
mind the older disks, but having that big beast sucking down power is a
no-go.

   I'd like to put a few drives in it and make it a network share for
the wife's Media PC and my Macs. But I just refuse to leave it running
all the time.

   Can I put it in a smaller case? Or run another OS with better power
management? Or am I really better served tossing it?

   Thanks for any suggestions from people who've gone through this
thought process before.

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

[Joshua Boyd]

On Mon, May 05, 2008 at 12:41:28PM -0400, William Enestvedt wrote:

>     There's a big ol' tower PC (with WinXP) in my basement I would like
> to throw out -- well, I just want to be rid of the bag power draw,
> really. The slower (c. 800 MHz?) CPU doesn't bug me much, and I don't
> mind the older disks, but having that big beast sucking down power is a
> no-go.

>    I'd like to put a few drives in it and make it a network share for
> the wife's Media PC and my Macs. But I just refuse to leave it running
> all the time.
>
>    Can I put it in a smaller case? Or run another OS with better power
> management? Or am I really better served tossing it?

I could be mistaken, but it seems to me that new drives might very well
use less power.

I doubt that case size makes any difference.  A new PSU could make a
difference, if you bought a high efficiency model.

Anyway, have you looked at any of the various off the self linux NAS
devices?  There are several fairly affordable models that shouldn't take
much power, and should support GigE nicely.  Presumably some of them
officially support installable software, but most of them are pretty
trivial to get root on to add your own stuff to them.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[William Enestvedt]

Josh wrote:
>
> I could be mistaken, but it seems to me that new drives might
> very well use less power.
>
   Fair point. But what I have are several IDE drives and this PC. :7)
But that leads into this:
>
> Anyway, have you looked at any of the various off the self linux
> NAS devices?  There are several fairly affordable models that
> shouldn't take much power, and should support GigE nicely.
>
   Bluntly, I haven't looked much past the prices I saw, most north of
$200, and I'm operating on a budget of under $100 plus any parts I can
scrounge at home or here at work. So I gotta ask, what's "affordable" to
you? (Knowing something about your life changes from the past few years
[which reminds me: did I see you write "wife"? Congratulations!], I
think I'm where you were financially, about a year ago -- only with me
it's the latest kid.)

   I like the idea of flashing one of these "simple" devices and turning
loose more of the Linux inside. I have seen the Linksys devices (too
big, too expensiove), and I saw the now-cancelled "Slug" (SLU-2?). What
else is out there that you had in mind when you wrote? Beside giving
"cheap nas linux" to Google, I mean?

   I have an IPX plus a really nice external SCSI case (conatining and
IDE-to-SCSI converter board), which is my last resort. *gulp*

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

[Mark Benson-2]

On 5 May 2008, at 15:05, Shannon Hendrix wrote:

> On May 5, 2008, at 00:50 , Geoffrey S. Mendelson wrote:
>
>> On Sun, May 04, 2008 at 08:41:19PM -0400, Sridhar Ayengar wrote:
>>> We are still around, but many of us have moved onto other things.
>>> Lately I've been writing custom financial applications in Java.
>>
>> Honestly, how much does knowing how the "bare metal" works help you
>> in writing code that runs in a virtual machine, on a different  
>> computer?
>
> A tremendous amount.

I'm a newcomer to coding (I writing PHP which is about as far detached  
from a real computer as you could want to be) and self taught but even  
I find knowing the underlying programming principles don't vary. Still  
even at high levels the core understanding of variables, functions,  
loops, arrays, test cases and the like all form an essential part. In  
effect most languages are formed around the same rules, but the higher  
you go the more abstract they are and the more wordy they become.

Knowing how a computer thinks is often invaluable. Especially when you  
assume how clever something in a high level language is, forgetting  
that underlying it is still the same old thinking and decision making  
engine that you use at assembler level.

For those who get uppity about Comp Sci degrees all being based around  
high level language, I think you have the picture a little skewed,  
even if you are correct. I majored in Geology, but my flat mates all  
did Comp Sci (and a good bit of it did rub off on me thankfully)  
Although the majority of programming exercises were high level  
language, usually in C or Java, they did teach a great deal about the  
basic workings of the machine itself and the principles of every layer  
of operations. The idea is Java is quick to learn and easy to grasp,  
an objet oriented (which like it or not is becoming very important).  
It also teaches general principles adequately that students can then  
extend their knowledge as the course goes on by looking into other  
systems if they feel the need.

Of course, I went to a good University (I'm not just saying that off-
hand, Southampton (UK) is a very high grade University according to  
the level of teaching and the quality of the graduates) and they are  
one of the UK's (if not the world's) premier Computer Science  
education and research establishments. Other Comp Sci degrees may  
vary :)

I don't profess to be an expert on anything, let alone Comp Sci, but  
comparing what I saw there to what you guys are talking about I can  
see they seem like they do a good job. Just my 2-peneth.


--
Mark Benson

My Blog:
<http://markbenson.org/blog>
Visit my Homepage: <http://homepage.mac.com/markbenson>

"Never send a human to do a machine's job..."
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Mon, May 05, 2008 at 02:28:20PM -0400, William Enestvedt wrote:

>    Fair point. But what I have are several IDE drives and this PC. :7)
> But that leads into this:

How many ide drives are you actually trying to use?

> >
> > Anyway, have you looked at any of the various off the self linux
> > NAS devices?  There are several fairly affordable models that
> > shouldn't take much power, and should support GigE nicely.
> >
>    Bluntly, I haven't looked much past the prices I saw, most north of
> $200, and I'm operating on a budget of under $100 plus any parts I can
> scrounge at home or here at work. So I gotta ask, what's "affordable" to
> you? (Knowing something about your life changes from the past few years
> [which reminds me: did I see you write "wife"? Congratulations!], I
> think I'm where you were financially, about a year ago -- only with me
> it's the latest kid.)

Yes, the 5 year aniversary is coming up in under 2 weeks.

Affordable obviously varies.

BTW, while I too am looking to do a new home server in the near future,
I don't want it to be an appliance, I don't want it to be Linux, and I
don't actually want to spend $600.  Perhaps $400, hopefully less.

>    I like the idea of flashing one of these "simple" devices and turning
> loose more of the Linux inside. I have seen the Linksys devices (too
> big, too expensiove), and I saw the now-cancelled "Slug" (SLU-2?). What
> else is out there that you had in mind when you wrote? Beside giving
> "cheap nas linux" to Google, I mean?
>
>    I have an IPX plus a really nice external SCSI case (conatining and
> IDE-to-SCSI converter board), which is my last resort. *gulp*

I don't recommend the IPX idea.

I'm assuming that you are using a P3, based on the mention of 800mhz
CPU.  Wikipedia suggests that that CPU is somewhere around 25 watts,
which isn't too bad (assuming it only a single CPU system).  

Since you mention older IDE drives, I would recommend getting that
down to 1 OS drive and 1 data drive or 2 data drives mirrored.  You
don't mention you target amount of disk space.  I would assume it is as
much as possible.  

I would recommend that you use whatever seems to be the coolest IDE
drive for the OS drive.  I would then recommend that you start with a
PCI SATA controller.  That should be about $15.   Since you mention the
IPX, I'm assuming that one data drive is OK with you.  So, spend another
$60 or so on a 250 gig SATA drive for sata.  

After that, my next consideration (but this would cause you to go past
$100) would be a new low power motherboard and a high efficiency PSU.  I
would imagine the pair would be about $150, and I think that HDs are
likely to be the low hanging fruit.

I'm thinking to use Solaris/SPARC with mirrored SCSI boot drives and
mirrored SATA drives (on a SAS controller) for data.  I have one SATA
drive and the gige card so far.  I need to get around to finishing
purchasing the rest of the parts.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Mon, May 05, 2008 at 07:49:13PM +0100, Mark wrote:

> I'm a newcomer to coding (I writing PHP which is about as far detached  
> from a real computer as you could want to be) and self taught but even  
> I find knowing the underlying programming principles don't vary. Still  
> even at high levels the core understanding of variables, functions,  
> loops, arrays, test cases and the like all form an essential part. In  
> effect most languages are formed around the same rules, but the higher  
> you go the more abstract they are and the more wordy they become.

I personally would rather use a language that is a good bit further from
the real computer than PHP is.  

But I still want to know how that language maps to the bare hardware.
 
> Although the majority of programming exercises were high level  
> language, usually in C or Java, they did teach a great deal about the  
> basic workings of the machine itself and the principles of every layer  
> of operations. The idea is Java is quick to learn and easy to grasp,  
> an objet oriented (which like it or not is becoming very important).  

You say it is important, but all the cool kids are moving on to things
like function programming, flow programming, logic programming, or
aspect oriented programming.  ;)

The smart student would prefer a nice versitile language that offers all
of the above, plus OO and procedural, while also making it easy to write
new language extensions and your own compilers.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[William Enestvedt]

Dang, Josh, five years? I guess it's too late to send a set of
candlesticks...

   A funny thng about "home servers" is that we want server-grade gear
but as quiet as a home office PC -- or at least my wife does, :7)

   There's a few V100s or Ultra 10s around here, but I'm reluctant to
even scrounge one of those b/c she would ask why I was running something
24/7...And she's be right. Perhaps I will try an Ultra 10 and just boot
it up manually when I want data off it, and pray that it never died in
its sleep. :7) Or *forehead slap* maybe this PC would be fine with a BSD
on it if I only ran it when I needed it -- and in that case, I may even
find a use for the old Happauge TV card!

   I so seldom read posts on this list about turning off machines when
idle (vice that "The Terror of Machine Room Silence" thread I enjoy so
much last week) that I never think of it myself.

   Josh, if you're avoiding Linux and appliances, are you thinking ZFS
on Open Solaris/reeBSD?

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

[Joshua Boyd]

On Mon, May 05, 2008 at 03:33:17PM -0400, William Enestvedt wrote:
> Dang, Josh, five years? I guess it's too late to send a set of
> candlesticks...
>
>    A funny thng about "home servers" is that we want server-grade gear
> but as quiet as a home office PC -- or at least my wife does, :7)

My wife has given up on the silent.  At this point I'm bothered by it
more than she is.  I love her so much.

If you will allow a amusing (at least to me) story about her, here is
one.  For the first 4.5 years, we lived in small apartments.  Somewhere
in the first two months of marriage (we moved about 2.5 months in), I
aquired both a machine rack and an Octane(that was actually a wedding
gift.  The other geeky wedding gift was a TI89).  From that point until
we moved again in October, she has lived with both in her living room.
And then at some point, I added an E250 as well (that box may have been
a bit of a mistake).  In the second apartment, there was no place you
could go and not hear the machines.

So, when we moved, the rack and most of the machines went into the
basement.  Our bedroom and office are on the second floor.  For about
the first week she complained that she couldn't sleep because it was
too quiet.  When I got a Fuel (a very nice trade) and put it in the
office next to the bedroom, she commented that she appreciated the noise
(but it is off when not in use because I don't actually appreciate the
noise).

When we have people over, she will proudly point out the machines, and
other women would fairly often tell her that they could never put up
with such things.

Did I mention I really love my wife?  And its not like I married a
computer person either.
 
>    There's a few V100s or Ultra 10s around here, but I'm reluctant to
> even scrounge one of those b/c she would ask why I was running something
> 24/7...And she's be right. Perhaps I will try an Ultra 10 and just boot
> it up manually when I want data off it, and pray that it never died in
> its sleep. :7) Or *forehead slap* maybe this PC would be fine with a BSD
> on it if I only ran it when I needed it -- and in that case, I may even
> find a use for the old Happauge TV card!

If you can stand the boot times, that is probably best.
 
>    I so seldom read posts on this list about turning off machines when
> idle (vice that "The Terror of Machine Room Silence" thread I enjoy so
> much last week) that I never think of it myself.

I've been leaving machines off more.  There are 5 machines I refuse
to leave off though.  My desktop Mac, my file server, my router, my
network services machine, and my... lets call it a DMZ.  I don't have a
static IP, so there are no public internet services running from home.
I'm not the only one who can access my home network remotely though.

I would be happy to let the Mac sleep, except that sleep doesn't work
correctly (I think it is the SATA card that is the problem), and if it
did sleep, I don't know how to do wake on lan.

I could let the file server off when I'm not home, but it takes forever
to boot.  If it had a working sleep function, I'd be happy to have it
asleep when I'm not home.  And I suppose if I was smart, I could turn it
off in the morning, then turn it on via the rsc over the internet before
I leave work to go home.  BTW, I try to work off the file server a lot,
rather than just copy stuff to it from the desktop periodically.

I do contemplate someday replacing the firewall (currently a Netra T1)
and the network services box ( a SS20) with lower power machines.  The
network services could be linux I suppose, perhaps on a small arm
board.  I could stick a Cirrus Logic ARM9 in a rack case. The CL board
is 200mhz with 32+ megs of ram and 128+ megs of flash, for $65.  It only
takes 0.6A of 5V power.  That would come to under $1 a day to power.
Finding such a small board to run affordably run NetBSD (I love the PF
firewall system) for a router would be a bit trickier.  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).

>    Josh, if you're avoiding Linux and appliances, are you thinking ZFS
> on Open Solaris/reeBSD?

Yes.  ZFS on Solaris on SPARC.  I really love ZFS and Solaris.  I've
loved Solaris since the mid 90s when I was exposed to it as a middle
schooler, and ZFS is for general purposes superior to anything else in
my opinion.

I also like NetBSD, but I don't believe that the ZFS port is usable
yet.  

And I do like linux more than a little.

And OSX is amazing in many small ways.  I'd love for Linux (and here I
probably mean GNOME most of all) to figure out how to improve in a large
number of areas.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Mike Loewen]

On Mon, 5 May 2008, William Enestvedt wrote:

>   A funny thng about "home servers" is that we want server-grade gear
> but as quiet as a home office PC -- or at least my wife does, :7)
>
>   There's a few V100s or Ultra 10s around here, but I'm reluctant to
> even scrounge one of those b/c she would ask why I was running something
> 24/7...And she's be right. Perhaps I will try an Ultra 10 and just boot
> it up manually when I want data off it, and pray that it never died in
> its sleep. :7) Or *forehead slap* maybe this PC would be fine with a BSD
> on it if I only ran it when I needed it -- and in that case, I may even
> find a use for the old Happauge TV card!

    I'm running a aged Dell Precision Workstation 530, which is perfectly
happy as a home server.  Dual Xeon 2.4GHz running Linux, and amazingly
quiet.  It runs 24/7, and runs the phones (Asterisk), the firewall, and a
PVR with MythTV, plus shared filespace (NFS, Samba and Netatalk) for
systems around the house.  The fan on my Fore ES-2810 switch makes more
noise than the server.  Backups go to a Powervault 120T with a DLT-7000
inside.


Mike Loewen mloewen@...
Old Technology http://sturgeon.css.psu.edu/~mloewen/Oldtech/
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Joshua Boyd]

On Mon, May 05, 2008 at 02:00:03PM -0600, 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.

Depending on the type of performance needed and the sort of problem you
are working on, OO code may be a very good idea.

I say that it is best to start with the most programmer efficient method
that could possibly work, then optimize from there.  If the fastest way
to get up and running and to prove your idea is to use a HLL running
across a cluster, then do so first, and optimize from there.  And
depending on the HLL you choose, you may not have much more work to do.

I don't know how well it normally works, nor do I know how to use it at
all, but I've seen some very impressive parallel performance from Jocaml
in demonstrations.  And for other situations there is Erlang.  And even
Python has potential to be an acceptable starting point.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Robert Darlington]

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.

-Bob

On Mon, May 5, 2008 at 1:03 PM, Joshua Boyd <jdboyd@...> wrote:

> On Mon, May 05, 2008 at 07:49:13PM +0100, Mark wrote:
>
>  > I'm a newcomer to coding (I writing PHP which is about as far detached
>  > from a real computer as you could want to be) and self taught but even
>  > I find knowing the underlying programming principles don't vary. Still
>  > even at high levels the core understanding of variables, functions,
>  > loops, arrays, test cases and the like all form an essential part. In
>  > effect most languages are formed around the same rules, but the higher
>  > you go the more abstract they are and the more wordy they become.
>
>  I personally would rather use a language that is a good bit further from
>  the real computer than PHP is.
>
>  But I still want to know how that language maps to the bare hardware.
>
>
>  > Although the majority of programming exercises were high level
>  > language, usually in C or Java, they did teach a great deal about the
>  > basic workings of the machine itself and the principles of every layer
>  > of operations. The idea is Java is quick to learn and easy to grasp,
>  > an objet oriented (which like it or not is becoming very important).
>
>  You say it is important, but all the cool kids are moving on to things
>  like function programming, flow programming, logic programming, or
>  aspect oriented programming.  ;)
>
>  The smart student would prefer a nice versitile language that offers all
>  of the above, plus OO and procedural, while also making it easy to write
>  new language extensions and your own compilers.
>
>
> _______________________________________________
>  rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

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

[Ethan O'Toole]

> 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.

It is not a screen scrape.

Anything that writes directly to the video card to speed up display will
bypass PC serial consoles. 3ware RAID card setup screens do this. Linux
kernel does as well, but you can obviously set it to the same serial port
and it will redirect.

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.

IPMI 2.0 can do video console redirection over ethernet for around $100 a
box (on those that support it). Compaq has their own stuff with ILOM. And
you can do remote power shutdowns via command line with IPMI. Both
controlled and pysical power off. And you can bring them back on via IPMI
as well. It's like LOM via ethernet.



--
05 REM Signature
10 PRINT " Ethan O'Toole "
20 PRINT " FLICKR ", " http://www.flickr.com/photos/ethanotoole "
30 PRINT " YOUTUBE ", " www.youtube.com/user/telmnstr "
40 PRINT " HOMEPAGE ", " users.757.org/~ethan "
RUN
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Brock Johnson]

Joshua Boyd wrote:

> I do contemplate someday replacing the firewall (currently a Netra T1)
> and the network services box ( a SS20) with lower power machines.  The
> network services could be linux I suppose, perhaps on a small arm
> board.  I could stick a Cirrus Logic ARM9 in a rack case. The CL board
> is 200mhz with 32+ megs of ram and 128+ megs of flash, for $65.  It only
> takes 0.6A of 5V power.  That would come to under $1 a day to power.
> Finding such a small board to run affordably run NetBSD (I love the PF
> firewall system) for a router would be a bit trickier.  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).
>  

NetBSD runs fine on the PCEngines ALIX boards,
http://www.pcengines.ch/alix.htm, while they're not quite as cheap as
that board you're quoting (about double the cost actually), they are
much better spec as well, 500mhz geode/256 megs of ram for the bigger
configuration. Use about 0.5a at 12v (to quote the manufacturers usage
rating, I haven't gotten the ammeter out and measured mine yet), so the
power is quite similar. I've got an 3c1 at the house I've been playing
around with as a wireless bridge and I'm about to get another model (the
2c3) to replace my old Soekris 4501.

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

[Joshua Boyd]

On Mon, May 05, 2008 at 03:29:36PM -0500, Brock Johnson wrote:

> NetBSD runs fine on the PCEngines ALIX boards,
> http://www.pcengines.ch/alix.htm, while they're not quite as cheap as
> that board you're quoting (about double the cost actually), they are
> much better spec as well, 500mhz geode/256 megs of ram for the bigger
> configuration. Use about 0.5a at 12v (to quote the manufacturers usage
> rating, I haven't gotten the ammeter out and measured mine yet), so the
> power is quite similar. I've got an 3c1 at the house I've been playing
> around with as a wireless bridge and I'm about to get another model (the
> 2c3) to replace my old Soekris 4501.

The 2c1 and 2c3 both look good.  I like the extra serial ports.  And the
I2C port is also eye catching for setting up a control panel.

I wonder if I can easily get a 44pin->40pin IDE adapter to put a CD-R on
that board.  If so, that might be perfect for another project I have on
the backburner.
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Jonathan C. Patschke]

On Mon, 5 May 2008, Joshua Boyd wrote:

> You say it is important, but all the cool kids are moving on to things
> like function programming, flow programming, logic programming, or
> aspect oriented programming.  ;)
>
> The smart student would prefer a nice versitile language that offers
> all of the above, plus OO and procedural, while also making it easy to
> write new language extensions and your own compilers.

And this is the perfect example of why computer science folks and
computer engineering folks are not alike.  Tools like that are fun and
expressive, and they tend to be quite excellent for -modeling- the
problem to be solved.  However, applying that sort of thinking to
deliverable software is what gives us what should be small applications
(web browsers, mail clients, IM clients), but end up taking hundreds of
megabytes of memory in their efforts to implement specific needs as
special cases of far more generic problems.  It's also the reason why
software tends to get slower[0] than faster.

Sometimes you just don't have the luxury.  Sometimes you just have to
get something done with some truly atrocious tools that meet some
reasonable intersection of usable and fast (leaning more towards the
latter).  And sometimes[1], programmer time is not more valuable than
run-time.

Take, for example, one of my current projects, which is a pile of code
that gets run against every part that $ork[2] ships out.  It's already
written in lean C, but I've reimplemented part of it in a special-case
scripting language (and written a compiler that generates PA-RISC
assembly code) because the ancient[3] version of HP's compiler we're using
generates such awful code.  Shaving a second off the runtime means being
able to run thousands more parts through the test facility in a day.
Sacrificing memory to generate full-gamut lookup tables for certain hash
functions (to save the hash computation at each dispatch) gave us a few
seconds of improvement.

Something like Python or Perl or Ruby[4] would cost us hundreds of
thousands of dollars a day (if not more) in tester execution time, and
would either require us to lower our production projections or purchase
more testers and hire more technicians to run them and buy more
buildings to put them in and etc.

Now, you're surely thinking this is a special case, right?

I look at it this way:  If I'm all about more money, I can get it.  I
might have to break laws or hurt people, but I can get it.  If I'm all
about getting more tail, the same thing applies.  If I'm all about any
material thing or experience, the same thing applies.  There's one
exception though:  I can't get more time.  Once my time on this planet
is up, that's it, as far as science can prove.

And every time my computer wastes my time because some retard programmer
got into a "look who's the cleverest" contest with some other retard
programmer, I take personal offense.  I did not buy the tools on my desk
for the amusement of others.  I bought it to make large piles of work
turn into large piles of deliverables.


[0] Try using a Mac SE/30 with software from the period when it was
     released.  Then switch to a modern Mac (or modern PC) and use
     today's software.  Nearly everything takes longer to launch, and
     isn't quite as snappy.  Why not?  The computer on my desk has over
     ONE THOUSAND times as much memory, four times as many CPUs, each of
     which is clocked just under TWO HUNDRED times as fast, nevermind the
     improvements in memory access times.
[1] In this case, I mean "usually".
[2] $ork builds integrated circuits.
[3] Unfortunately, upgrading is not reasonable.  Nor is switching to
     gcc.  We're using what is certified with the IC tester, and as
     broken as the compiler is, it's entirely possible that the tester
     interface software expects some of that brokenness.
[4] Not to say that they haven't their place.  Another project I'm
     working on actually uses ECMAscript as its embedded scripting
     language, specifically because it has closures and prototype-style
     OO.  However, in that project, the scripting engine does not run
     synchronously with the test payload.
--
Jonathan Patschke | "There is more to life than increasing its speed."
Elgin, TX         |                                   --Mahatma Gandhi
USA               |
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue

[Lionel Peterson-2]

On May 5, 2008, at 3:46 PM, Joshua Boyd <jdboyd@...> wrote:

<snip>

  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 w