PCB CAD
Don Prescott
and shareware, give me one well known organisation using "open source"
PCB layout tools on mission critical projects...?
I'm waiting.............!
Prescott
Stuart Brorson
: OK.... after all that wuffle about the distinction twix open source
: and shareware, give me one well known organisation using "open source"
: PCB layout tools on mission critical projects...?
: I'm waiting.............!
I apologize for not responding sooner; I have been traveling, and away
from the net for a while.
Asking about EDA for "mission-critical" applications make you sound
like a marketing flack for one of the various EDA companies. Are you
trying to spread a little FUD by implying that free/open-source tools
aren't up to the hypothetically exacting demands of a Fortune-500
company working on massive projects, and that you can therefore
dismiss them out of hand?
*chuckle*
Let's be clear: What is the current market segment served by gEDA,
PCB, Icarus, and other F/OSS EDA projects? These tools are written
*for*, and used *by* folks at the low and middle segments of the CAD
marketplace. They are ideal for students, hobbiests, educators, and
small-time consultants -- a market segment which has been badly served
by the EDA vendors (who offer bad service, ridiculously high prices,
and buggy products). Right now, F/OSS tools are workable replacements
for things like Circuit Maker and Electronics Workbench. Indeed, if
you want to do a multi-page design with little or no hierarchy, and
lay it out as a 4 or 6 layer board, you might have previously used one
of these apps, or even Orcad or Protel. Now, for this kind of job
you can use gEDA instead.
As for "mission critical" designs: Do Fortune 500 companies use
Circuit Maker or EWB for "mission critical" designs? No way. Is
Protel capable of "mission critical" projects? That's debatable at
best -- my experience with it was that it was buggy and flakey, but
with enough swearing & futzing around you could ultimately get the job
done anyway.
Therefore, your question of whether the F/OSS apps are ready for
"mission-critical" design is a red herring; none of the apps which
serve the low to middle market segment are really "mission critical".
Note that with the gEDA suite (which is generally less buggy than
Protel, by the way), you can not only get the job done, but the
program is free, offers ASCII file formats, legacy design protection,
and is unencumbered by licencing restrictions. *That's* the important
take-away point.
Now, are people using open-source EDA for low and mid-level projects?
Yes. Here are some public examples:
Ronja free space optical link:
http://ronja.twibright.com/
D. Harmon's Single Board Computer:
http://www.dlharmon.com/sbc.html
Mikey Sklar's electric clothing:
http://www.electric-clothing.com/tools.html
ESNUG -- DeepChip website:
http://www.deepchip.com/items/dvcon04-02.html
(about 1/2 way down there is a mention of Icarus Verilog used at
InformASIC.)
These are the publically announced projects which I am personally
aware of without doing a Google search. If you Google around, you
will find others. Also, I am aware of some people using gEDA & Icarus
in "stealth mode" inside larger corporations. However, it's not up to
me to give away their identity, so I am sorry that I cannot provide
any backing to this assertion. I realize this is unsatisfying.
-----------------------------------------------------------------
That's the present. What about the future? Consider again the
advantages offered by gEDA and other F/OSS apps:
* Open-source.
* Documented ASCII file formats.
* Legacy design protection.
* They don't tie you up in licencing knots.
* Download is free.
These advantages are important to practicing engineers, but are
antithetical to the interests of commercial EDA vendors, who need to
find ways to lock down their customers and then squeeze money from
them by every device available. Consequently, it is unlikely that the
commercial EDA vendors will be able to offer the same features.
Therefore, they can't compete against the F/OSS stuff in the long run,
at least on these parameters.
Meanwhile, as more and more people become aware of F/OSS EDA projects
like gEDA, the projects will acquire developers, who will fix more
bugs, introduce more features, create more apps, and so on. So, on
one hand, the commercial vendors have frozen development of their
low-end stuff (because low-end CAD is commoditized & doesn't make
money), and will not meaningfully implement any of the above features
which are important to engineers. On the other hand, the F/OSS stuff
will continue to grow. I think you can see where it will end up.
At the high end, however, it's another story. At the high end, the
EDA vendors & VCs interested in EDA have lots of money to
lavish on research & development of new tools and algorithms. The
F/OSS guys don't have the time, resources, or interest to compete at
this level. Therefore, I don't expect to see people using PCB instead
of Allegro or Mentor Expedition any time soon. However, for low and
mid-level size designs -- the bread and butter of students, hobbiests,
and small consultants -- I fully expect to see gEDA and other F/OSS
applications take over the market place over the next decade, simply
because the commercial EDA vendors won't & can't compete.
Stuart
David
It's interesting to compare the EDA market with the embedded software
development tools market, which is closely related. The difference is
that there is not so much in the way of seriously high-end, high-price
software development tools (you would be hard pushed to spend a budget of
$20,000 for one target - except perhaps for some emulator hardware), and
there are more users, especially at the low end. The market is also a bit
more mature in many ways (perhaps because of the larger number of users),
and I think might give an indication of the future in the EDA market.
In particular, there are high-end tools, with compilers from Green Hills,
Diab Data, Metrowerks and the like - these are the best
compilers/debuggers that money can buy, and clearly outclass the cheap and
free tools. These vendors have learned to live with open-source software,
and compete against it on quality, support and features, combined with
supporting it (such as supporting embedded linux targets). At the
mid-range, companies like IAR have a certain stable base (their key
selling point is a wide range of support for targets), but I don't think
their future is solid. A few years ago, microcontroller vendors would
work closely with IAR on new architectures, and support them as their
"official" compiler - the market is now rebelling against that, starting
from the low-end but working up. Companies like Atmel and Texas
Instruments listen when users say they would buy their chips if they had a
wider range of tools available, and you now see links to gcc ports on
vendors web sites. In addition to the growing inroads made by open source
software here, there is also a growth market in "cheap and cheerful"
tools. These vendors specialise in making software that is easy to use
and providing great support. Typically they are low on features, and
technically mediocore (in terms of optimisations, for example) - but that
doesn't matter much to their market. I can imagine this sort of market
growing significantly in the EDA world too.
There is another big advantage of open source tools - they are useful for
other vendors. As an example, look at Altera and their tools for FPGA
design and especially for their Nios II soft processor. Instead of
re-inventing the wheel, or buying in proprietry technology, they use tcl
and perl extensively for scripting. The do at least as good a job as any
proprietry scripting language, are well-known, and are free - important if
you want to be able to give away free, limited copies of your software.
When they made their Nios processor, instead of writing a compiler toolkit
from scratch, or working with a big commercial vendor, they ported gcc.
This gave them a solid basis to work with. When they wanted an IDE, the
started with eclipse and wrote some plug-ins. I can also imagine this
sort of thing becoming more common in the EDA world - if I wanted to sell
a new simulator program, I'd bundle it with gEDA to form a complete
package. Sure, I'd make sure it would work nicely with Protel and Mentor
and all the other major commercial packages - but it would give me a head
start for little effort.
Don Prescott
These tools are written
> *for*, and used *by* folks at the low and middle segments of the CAD
> marketplace. They are ideal for students, hobbiests, educators, and
> small-time consultants
You wrote an awful lotta stuff, but then agreed with what I said...
YES, open source CAD applications are probably OK for "students,
hobbiests, educators, and small-time consultants". Generally
speaking, nobody uses these sort of products when you need to get a
commercial product out the door.....(mission critical). Why
not...???? 'cos you're much more concerned with the $$$KKK than
playing around with some home-grown piece of code...
Oh.... please don't tell me about apache and linux (again), thems was
never what I was talking about...
Prescott
David
You are the one who has been generalising about all open source software
being inferior to commercial products, with comments such as "Who in there
[sic] right mind would consider using an open source wordprocessing
package or spreadsheets instead of the commercial products like Word and
Excel". As long as you continue to mix in such absurd claims, you'll get
appropriate counter-claims.
There are relatively few people who have the money to buy top
commercial EDA applications and yet choose to use open source software -
no one is disputing that. But there are a few such users, and the base is
growing along with the quality and usefulness of gEDA and friends.
Projects such as linux, apache and open office make it perfectly clear
that, in general, open source software can be at least as good (and often
far better) than commercial software. Successful companies like MySQL
demonstrate that it is possible to make a good business from making and
selling open source software. How successful open source software will be
in the EDA market remains to be seen - it is likely to make increasing
inroads at the low end, with little effect at the high end.
> Prescott
in...@bostonsemiconductor.com
mission critical development?
Has anyone ever used emacs to write source code to tape out a chip, or
write a line of C++ or assembler for a shipped executable? And how
would that C++ have been compiled, nobody uses a GNU compiler right?
How many boards, chips, executables can anyone really ship without
touching an open source tool along the way? I bet there aren't that
many.
Check out the Altera website. They have 14,000 (paying) customers and
made over $800 million last year. Or the Xilinx website. A swarthy
group of "mission critical" customers gave them 1.4 billion last year.
Each of them _ship_ cygwin and GCC as part of their development flow.
Last I checked, those tools are open source.
I have an idea - just the one, I'm good for one a day and this is it.
Rather than throwing rocks, why don't the folks who think closed source
is a good idea try rationalizing it. Make a case for it - what are
it's inherit benefits? How are customers better off when they use it?
Stop making a case against open source development - it's not worth
your time, right? Start making a case _for_ closed source.
If we all listened when someone said, "you can't do that", we'd never
do anything.
Chris
Don Prescott
and
>made over $800 million last year. Or the Xilinx website. A swarthy
>group of "mission critical" customers gave them 1.4 billion last
year.
>Each of them _ship_ cygwin and GCC as part of their development flow.
>Last I checked, those tools are open source.
The theme is PCB CAD matey – YES, I know all about FPGA design
tools... Altera and Xilinx's interest is in selling FPGAs NOT in
selling tools. They give away the design tools so you buy their
hardware. This wasn't what we were arguing about... Show me a well
known company using open source PCB CAD tools on mainstream $$$$KKKK
projects....not file servers! not FPGA! PCB CAD
Prescott
in...@bostonsemiconductor.com
tools.
I heard you the first 10 times when you said, "open source PCB sucks".
Now let's hear from you why close source PCB CAD is so great.
Chris
Don Prescott wrote:
> >Check out the Altera website. They have 14,000 (paying) customers
> and
> >made over $800 million last year. Or the Xilinx website. A swarthy
> >group of "mission critical" customers gave them 1.4 billion last
> year.
> >Each of them _ship_ cygwin and GCC as part of their development
flow.
> >Last I checked, those tools are open source.
>
> The theme is PCB CAD matey - YES, I know all about FPGA design
Don Prescott
>tools. I heard you the first 10 times when you said, "open source PCB
sucks".
>Now let's hear from you why close source PCB CAD is so great.
In any of the messages have you seen me use the term "open source PCB
sucks"....?
What I am saying is that in commercial organisations the EDA tools are
a means to an end only; a link in the chain leading to the final
production product. These organisations buy products that are ready
to use, reliable, and are supported. How to you think Cadence brings
in $1billion a year...? A hard-node, skunkwork, Engineering VP is
unlikely to be interested in considering an open source solution if a
commercial product is available. AND they are available at all sorts
of prices. Nowadays, quality PCB CAD products can be purchased for
several hundred dollars.
If you say I'm wrong then OK show me well known companies that do use
open source PCB CAD solutions for their mainstream PCB design
operation. AND PLEASE don't come back with words or phrases I haven't
used, or refer to non-PCB CAD apps like FPGA or file servers.
Prescott
David
[OT Question] Why do you keep starting new threads, rather than continuing
posting in the same thread?
>>I'd rather hear you make a compelling case for using closed source
>>tools. I heard you the first 10 times when you said, "open source PCB
> sucks".
>>Now let's hear from you why close source PCB CAD is so great.
>
> In any of the messages have you seen me use the term "open source PCB
> sucks"....?
Several of your messages could be summed up as saying "all open source
software sucks, and no commercial organisation would use them for
something important, especially not pcb design".
>
> What I am saying is that in commercial organisations the EDA tools are
> a means to an end only; a link in the chain leading to the final
> production product. These organisations buy products that are ready
> to use, reliable, and are supported. How to you think Cadence brings
If you are happy now to retract your previous claims regarding open source
software in general, and concentrate on specifically pcb design software,
then that's fair enough. As someone who does schematic design, pcb
design, fpga development and embedded systems programming, I have a
tendancy to lump together a lot more under the heading "EDA" than just pcb
design.
> in $1billion a year...? A hard-node, skunkwork, Engineering VP is
> unlikely to be interested in considering an open source solution if a
> commercial product is available. AND they are available at all sorts
> of prices. Nowadays, quality PCB CAD products can be purchased for
> several hundred dollars.
Only a half-wit, PHB would not consider using an open source solution. I
think in the case of pcb design software, as it stands today, few
(outside the very lowest budget users, or those with particular
requirements for openness) would choose it simply because there are
commercial packages that provide a better solution for most uses. In the
future, as the open source packages develope, then who knows which will
make the most sense?
My point is not that there are open source pcb design tools available
today that are a good choice for a commercial organisation - they are not
currently suitable for my own uses, and I doubt if they are suitable for
more than a small percentage of users. It is merely that the idea of
dismissing open source software out of hand simply because there are
commercial alternatives is absurd - it's an old-fashioned and out of date
prejudice that is being repeatedly proven unwise in other software fields.
in...@bostonsemiconductor.com
unfair treatment to say that your posts suggested you thought Open
Source PCB software "sucks".
I am glad that you do not think it "sucks", however let me point out
some of the things you said, verbatim that IMHO indicate to the broader
audience that you think it does "suck":
-"raw, unsupported, buggy, pieces of software "
-"Useless if you're trying to design a real product that's gotta go out
the door on time"
You are correct. You did not say it sucks.
My bad.
Chris
Don Prescott
continuing
>posting in the same thread?
Cos' Google says it can't retrieve the original message
>Several of your messages could be summed up as saying "all open
source
>software sucks, and no commercial organisation would use them for
>something important, especially not pcb design".
You're twisting my words and adding your own .....yet again!! I don't
use or like phrases that have a sexual connotations like "sucks", so
PLEASE don't attribute them,in any way, to me...
>If you are happy now to retract your previous claims regarding open
source
>software in general, and concentrate on specifically pcb design
software,
>then that's fair enough. As someone who does schematic design, pcb
>design, fpga development and embedded systems programming, I have a
>tendancy to lump together a lot more under the heading "EDA" than
just pcb
>design.
What is the subject called..."PCB CAD"....! What is this special
interest group...... "electronics CAD"...Is it about servers and apps
in general...No!
>Only a half-wit, PHB would not consider using an open source
solution. I
>think in the case of pcb design software, as it stands today, few
>(outside the very lowest budget users, or those with particular
>requirements for openness) would choose it simply because there are
>commercial packages that provide a better solution for most uses. In
the
>future, as the open source packages develope, then who knows which
will
>make the most sense?
Dunno what PHB means.... but at least you agree with what I've been
saying...
>My point is not that there are open source pcb design tools available
>today that are a good choice for a commercial organisation - they are
not
>currently suitable for my own uses, and I doubt if they are suitable
for
>more than a small percentage of users. It is merely that the idea of
>dismissing open source software out of hand simply because there are
>commercial alternatives is absurd - it's an old-fashioned and out of
date
>prejudice that is being repeatedly proven unwise in other software
fields.
In conclusion, I very much doubt that open source PCB CAD will be of
interest to many commercial operations for a considerable time yet.
Prescott
in...@bostonsemiconductor.com
don't
> use or like phrases that have a sexual connotations like "sucks", so
> PLEASE don't attribute them,in any way, to me...
Cute. You're right, you never used sexual innuendo in your words here:
-"raw, unsupported, buggy, pieces of software "
-"Useless if you're trying to design a real product that's gotta go out
the door on time"
For the record, I intended no sexual connotation by the remark "sucks".
If you choose to view it that way, that is your right.
For the time being, I agree to disagree - I'm not even sure what you're
arguing about any more.
David
>>OT Question] Why do you keep starting new threads, rather than
> continuing
>>posting in the same thread?
>
> Cos' Google says it can't retrieve the original message
Strange... I haven't used google for posting, so I wouldn't know. Isn't
it easier to use a newsreader program?
>
>>Several of your messages could be summed up as saying "all open
> source
>>software sucks, and no commercial organisation would use them for
>>something important, especially not pcb design".
>
> You're twisting my words and adding your own .....yet again!! I don't
> use or like phrases that have a sexual connotations like "sucks", so
> PLEASE don't attribute them,in any way, to me...
Ok - I'm happy to agree that you never used such a word. But most people
here don't read "sucks" as having any sexual connotation, but is simply an
insulting way of saying something is bad - much like calling software
"crashware", and other choice phrases that you *did* use.
>
>>If you are happy now to retract your previous claims regarding open
> source
>>software in general, and concentrate on specifically pcb design
> software,
>>then that's fair enough. As someone who does schematic design, pcb
>>design, fpga development and embedded systems programming, I have a
>>tendancy to lump together a lot more under the heading "EDA" than
> just pcb
>>design.
>
> What is the subject called..."PCB CAD"....! What is this special
> interest group...... "electronics CAD"...Is it about servers and apps
> in general...No!
>
I agree - but other apps turned up in the context of your claiming that
all open source software was inferior, so counter-examples are perfectly
valid.
>>Only a half-wit, PHB would not consider using an open source
> solution. I
>>think in the case of pcb design software, as it stands today, few
>>(outside the very lowest budget users, or those with particular
>>requirements for openness) would choose it simply because there are
>>commercial packages that provide a better solution for most uses. In
> the
>>future, as the open source packages develope, then who knows which
> will
>>make the most sense?
>
> Dunno what PHB means.... but at least you agree with what I've been
> saying...
"Pointy Hair Boss" - are you an engineer who doesn't read Dilbert?? A PHB
is a manager with no understanding of what he is managing, and typically
believes ideas such as that you have to pay for software so that you have
someone to sue if it doesn't work (have you ever read a software license?
You can sue the supplier if the CD is scratched or the pages fall out of
the manual within 3 months, and that's about it).
>
>>My point is not that there are open source pcb design tools available
>>today that are a good choice for a commercial organisation - they are
> not
>>currently suitable for my own uses, and I doubt if they are suitable
> for
>>more than a small percentage of users. It is merely that the idea of
>>dismissing open source software out of hand simply because there are
>>commercial alternatives is absurd - it's an old-fashioned and out of
> date
>>prejudice that is being repeatedly proven unwise in other software
> fields.
>
> In conclusion, I very much doubt that open source PCB CAD will be of
> interest to many commercial operations for a considerable time yet.
I agree mostly, but probably for a shorter interpretation of "considerable".
>
> Prescott
Tim
I know of one aerospace company currently developing their next range of boards
using FreePCB. They key for them is that they can get the source, the fact that
it is also free is immaterial, though naturally they are delighted it's free, if
a little incredulous. Now that their eyes have been opened to open-source, a
snowball effect is in motion, they are very keen to get hold of open-source DSP
software.
I suspect that at the current moment in time there are very few companies using
open source PCB layout tools, though as in so many other areas of development,
there is an inexorable drift toward open-source software: it's the source that
counts.
Happy xmas, Tim.
Don Prescott
Re:
>I know of one aerospace company currently developing their next range
of >boards using FreePCB. They key for them is that they can get the
source, the >fact that it is also free is immaterial, though naturally
they are delighted >it's free, if a little incredulous. Now that their
eyes have been opened to >open-source, a snowball effect is in motion,
they are very keen to get hold >of open-source DSP software.
>I suspect that at the current moment in time there are very few
companies >using open source PCB layout tools, though as in so many
other areas of >development,there is an inexorable drift toward
open-source software: it's >the source that counts.
Oh yes...... "you know an Aerospace company that uses open source PCB
CAD" do you..?
This has the hallmarks of an urban myth. The term "Aerospace company"
conjures images of maybe Boeing, Grumman, or perhaps Lockheed... and
in the next retelling it will become one of these organisations..
OK what Aerospace company..??
Prescott