New Software

[Stephan Rose]

Well seeing a lot of complaints here on this NG about various design
products out there, and also trial testing many commercial packages
ourselves and cursing within minutes....(still using 12yr old tango
from dos, how sad is that?)

We've decided to simply..make our own. Me being the guy who's writing
the software.

Now....for anyone...who wants a chance to influence something like
this from the start..here's your chance...other than of course my
boss' wishes and requests of features he'd like to see...

What would you all like to see? Any input is appreciated and will be
considered if possible.

Here's things I've got planned so far, though these are mostly just
software features....not any electronic/technical features..those I'll
start defining as I get to stages where I could implement them.

1. Complete integrated IDE. That's right...no 10 different
applications from DIFFERENT companies that work together on mondays
and hate each other on fridays. So schematic design, pcb design,
library mangement, etc. can all be done seamlessly from within the
same application.

2. Plugin capability, ability for anyone to write their own plugins if
they wish for exporters/importers, your own router, even define your
own document types. Basically the entire API will be available via
this mechanism.

3. All drawing functions are hardware accelerated, DirectX9+. So
expect fast, smooth editing with realtime results instead of the
inverted line stuff, etc.

4. Support for all various types of units (current mm, cm, inch, mil.
Need more? which ones?) all at 96-bit fixed floating point precision
(fixed FP means no precision losses or precision problems so common
with standard FP).

Currently I'm working on the core engine, core drawing functions, and
the schematic component editor to start with and move up from there.
Expecting to have a working alpha or even beta by summer of this year.

So as I'm currently working on the component editor (schematic symbols
only), any feature that everyone particularly would like but nobody
has?

Here's the list of in what order I plan on implementing what

1. Core engine and functions - Current stage
2. Schematic Component Editor - Current stage
3. Schematic Design
4. Pcb Component Editor
5. Pcb Design

Then..once all that works....everything and everyone is happy...

Auto router and simulation. Whever makes sense first at that time.

So there it is...now how about some input?

Stephan

 

[Stuart Brorson]

[ . . . description of EDA project snipped . . . .]
: Here's the list of in what order I plan on implementing what

: 1. Core engine and functions - Current stage
: 2. Schematic Component Editor - Current stage
: 3. Schematic Design
: 4. Pcb Component Editor
: 5. Pcb Design

: So there it is...now how about some input?

Sounds interesting. But why re-invent the wheel? If you want
schematic capture and netlisting, why not use gEDA?

http://www.geda.seul.org/

If you think it doesn't have enough features, then why not take the
existing code base and add them?

As for PCB layout, there is already an existing program: PCB:

http://pcb.sourceforge.net/

PCB's GUI is primitive. If you feel like hacking, why not take PCB's
innards and build them into a GUI compatible with the 21st century?

IMHO, there are too many 1/2 completed free software projects out
there. Why not take one that already exists and extend it? (Note
that gEDA is already 100% functional, but could always use new
features. . . .)

Stuart

 

[Stephan Rose]

[ . . . description of EDA project snipped . . . .]
: Here's the list of in what order I plan on implementing what

: 1. Core engine and functions - Current stage
: 2. Schematic Component Editor - Current stage
: 3. Schematic Design
: 4. Pcb Component Editor
: 5. Pcb Design

: So there it is...now how about some input?

Sounds interesting. But why re-invent the wheel? If you want
schematic capture and netlisting, why not use gEDA?

Because I'm getting paid to do this and make my living with it.

Plus this is planned to be a major commercial product for our company
once done so all the code should be our own really. Even if it means
reinventing some wheels.

Also, I'm writing this in C# and the .Net Framework and none of these
existing projects would be compatible with that.

I highly appreciate the input though, and will regardless check those
things out just to see what they do and how. Can only learn.

Thanks,

Stephan

 

[Stuart Brorson]

: On Tue, 03 Feb 2004 23:51:44 -0000, Stuart Brorson <[email protected]>
: wrote:
:>Sounds interesting. But why re-invent the wheel? If you want
:>schematic capture and netlisting, why not use gEDA?

: Because I'm getting paid to do this and make my living with it.

Fair enough.

: Plus this is planned to be a major commercial product for our company
: once done so all the code should be our own really. Even if it means
: reinventing some wheels.

OK, fine. I assume that you are targeting the low end, because there
are already enough tools at the high end (i.e. Viewdraw, Allegro,
Mentor's stuff, etc.). Also, the big players (e.g. Cadence) are
ababdoning the low end, leaving market opportunity.

If you ask me, MicroSim's schematic capture & layout tools were the
easiest and best tools to use for simple board design.
Even today, MicroSim's integration of SPICE analysis with schematic
capture hasn't been beat. Therefore, if you could make a product
behaving similarly -- but without all the d*mn bugs (lockups, hangs,
unexpected behaviors, etc.), you would have a winning product.

: Also, I'm writing this in C# and the .Net Framework and none of these
: existing projects would be compatible with that.

These are MicroSoft languages & environments. Others will differ, I
am sure, but I wouldn't be writing products for any MicroSoft OS right
now. Momentum is building for a mass defection to Linux on the
desktop, and I believe that engineers will lead the charge. I have
done it already, and I am not alone: I know several guys who have put
Linux machines on their desktops at work & use the machines for as
much engineering work as they can. (They still must use Windows for
mail & word processing, but that wil come to an end shortly. . . . .)
Therefore, a good low-end EDA package which runs natively on Linux
would be a good thing for me and for your company. FWIW:
MentorGraphics is targeting Linux as a platform for their stuff.

: I highly appreciate the input though, and will regardless check those
: things out just to see what they do and how. Can only learn.

Good Luck!

Stuart

 

[Stephan Rose]

: On Tue, 03 Feb 2004 23:51:44 -0000, Stuart Brorson <[email protected]>
: wrote:
:>Sounds interesting. But why re-invent the wheel? If you want
:>schematic capture and netlisting, why not use gEDA?

: Because I'm getting paid to do this and make my living with it.

Fair enough.

: Plus this is planned to be a major commercial product for our company
: once done so all the code should be our own really. Even if it means
: reinventing some wheels.

OK, fine. I assume that you are targeting the low end, because there
are already enough tools at the high end (i.e. Viewdraw, Allegro,
Mentor's stuff, etc.). Also, the big players (e.g. Cadence) are
ababdoning the low end, leaving market opportunity.

That is going to be our initial target. Going to expand towards the
high end as time goes on. But low end is our first and initial
priority.

If you ask me, MicroSim's schematic capture & layout tools were the
easiest and best tools to use for simple board design.
Even today, MicroSim's integration of SPICE analysis with schematic
capture hasn't been beat. Therefore, if you could make a product
behaving similarly -- but without all the d*mn bugs (lockups, hangs,
unexpected behaviors, etc.), you would have a winning product.

Hence why I'm using the language and envrioment I am using.

It's not quite as fast as C++, but in exchange it offers extreme
stability. Invalid memory accesses, memory leaks, etc. (and crashes
caused by that) are virtually impossible. No pointer mess either. It's
beautiful

: Also, I'm writing this in C# and the .Net Framework and none of these
: existing projects would be compatible with that.

These are MicroSoft languages & environments. Others will differ, I
am sure, but I wouldn't be writing products for any MicroSoft OS right
now. Momentum is building for a mass defection to Linux on the
desktop, and I believe that engineers will lead the charge. I have
done it already, and I am not alone: I know several guys who have put
Linux machines on their desktops at work & use the machines for as
much engineering work as they can. (They still must use Windows for
mail & word processing, but that wil come to an end shortly. . . . .)
Therefore, a good low-end EDA package which runs natively on Linux
would be a good thing for me and for your company. FWIW:
MentorGraphics is targeting Linux as a platform for their stuff.

Actually C# and the .Net Framework are designed by microsoft to be
cross-platform and easily portable. When compiled, the code generated
actually is not executable code, but intermediate byte code instead.
Hence, my software would run on any operating system and hardware
supporting the .net framework. The actual executable code is created
at run-time on the target platform the executable is run on.

There already is a .Net Framework implementation (called mono) and C#
compilers, etc. available for linux. And from the looks of it..they
support most everything already. Maybe I'll give it a try one of these
days and see what happens.

Though the rendering core is Managed DX to where there is currently no
port available. But....those are confined to a single DLL which I
could easily swap for a linux version equivalent.

: I highly appreciate the input though, and will regardless check those
: things out just to see what they do and how. Can only learn.

Good Luck!

Thank you

Stephan

 

[Kevin Aylward]

Thanks,

Well seeing a lot of complaints here on this NG about various design
products out there,

Engineers like to complain. Engineers *always* think they can do it
better. They wouldn't be engineers if that weren't the case. However,
the naive oes exclude time and money in their evaluations.

We've decided to simply..make our own. Me being the guy who's writing
the software.

Oh dear...the sucker..

Now....for anyone...who wants a chance to influence something like
this from the start..here's your chance...other than of course my
boss' wishes and requests of features he'd like to see...

All in no time with zero cost.

Ok... here is is. You won't like it though. Look, every one of us has
these grandiose ideas about making some sort of great, super duper
perfect product. It will *never* happen. This is the real world.

Personally, I thing you and your company are completely wasting your
time. It *will* be time and money directly down the toilet. There is way
to much competition in general purpose cad for a new entry to make it.
You wont take sales away from the existing customers. The only way you
will get marketing share is by giving it away.

You seem to have the daft and naive idea that people buy and use the
"best" products. They don't. Secondly, you product can't possible be the
"best" because everyone wants something *different*. Its impossible to
cover all bases, this is so, even if the software writer is an *expert*
in the actual *use* of such software

Because I'm getting paid to do this and make my living with it.

This means you have a fundamental issue with evaluating the business
sense of doing this product.

Plus this is planned to be a major commercial product for our company
once done so all the code should be our own really. Even if it means
reinventing some wheels.

I think you just have not done enough market research on this. Its
flooded. Realistically, there is *zero* chance of a new same shit,
different day cad product making it today.

Also, I'm writing this in C# and the .Net Framework and none of these
existing projects would be compatible with that.

I highly appreciate the input though,

I suspect you wont, as people generally only like positive comments,
they ignore the negative ones. People seeking advice are usually only
seeking an accomplice.

and will regardless check those
things out just to see what they do and how. Can only learn.

So, you *already* have decided on this idea of doing a product, that 30+
other companies already make, and yet haven't really seen what is
already out there and *why* they are being used.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

"That which is mostly observed, is that which replicates the most"
http://www.anasoft.co.uk/replicators/index.html

"quotes with no meaning, are meaningless" - Kevin Aylward.

 

[Kevin Aylward]

Even today, MicroSim's integration of SPICE analysis with schematic
capture hasn't been beat.

I have to disagree with this

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

"That which is mostly observed, is that which replicates the most"
http://www.anasoft.co.uk/replicators/index.html

"quotes with no meaning, are meaningless" - Kevin Aylward.

 

[Paul Camilleri]

The idea of rolling your own software is good. We reached the same
conclusion about a year and a half ago and started to write our own software
for schematic capture and simulation and PCB design. It is not intended as a
commericial application in it's own right but will be given away freely
along with a PIC C development environment to support a commercial hardware
product.

I agree about the mass migration to Linux (the more the merrier), but we
decided to take a multi-platform approach. Our software is written in Delphi
6.0/Kylix 2.0 for the frontend and C++ for stuff that is not OS specific.
I'm not trying to sell the toolset here but I cannot fault it for what we
are doing.

Good luck!

Regards

Paul Camilleri

 

[nyffeler]

Hi Stephan,

I would like to tell you my opinion of features required for a good E-CAD
software, without starting a long discussion in the newsgroup.
Would you mind to give me your e-mail address?

Peter

 

[Stuart Brorson]

: On Wed, 04 Feb 2004 02:23:01 -0000, Stuart Brorson <[email protected]>
: wrote:
:>OK, fine. I assume that you are targeting the low end, because there
:>are already enough tools at the high end (i.e. Viewdraw, Allegro,
:>Mentor's stuff, etc.). Also, the big players (e.g. Cadence) are
:>ababdoning the low end, leaving market opportunity.

: That is going to be our initial target. Going to expand towards the
: high end as time goes on. But low end is our first and initial
: priority.

I tend to agree with Kevin's point (elsewhere on this thread) that the
CAD market is flooded, particularly at the low end. YOu may have
trouble making money on this. That being said, I don't know of many
end-2-end suites which do schematic capture, SPICE simulation, *and*
decent, multi-layer layout. If I am missing one, please let me know.
For example, does Kevin's SPICE prog do layout?

An aside about your project: All file should be easily parsable
ASCII. Hiding things in binary files, or -- worse -- in "databases"
(a la Protel) is anathema.

: Hence why I'm using the language and envrioment I am using.

: It's not quite as fast as C++, but in exchange it offers extreme
: stability. Invalid memory accesses, memory leaks, etc. (and crashes
: caused by that) are virtually impossible. No pointer mess either. It's
: beautiful

Sounds like Java. . . .

: Actually C# and the .Net Framework are designed by microsoft to be
: cross-platform and easily portable. When compiled, the code generated
: actually is not executable code, but intermediate byte code instead.
: Hence, my software would run on any operating system and hardware
: supporting the .net framework. The actual executable code is created
: at run-time on the target platform the executable is run on.

Sounds like Java. . . . .

So what's to prevent it from being sloooooooooow, like Java? And
what's to prevent Microsoft lock-in, as C# & .nut morph ever so
slightly over time to keep everybody off balance unless they stick
with the latest version of Windoze? This is what has happened with
InternetExploiter and the various versions & file formats of
MSOffice. And this is also true of C# -- it's really just Microsoft's
attempt to derail Java by providing an equivalent language which only
Microsoft controls.

As a potential customer, I am already becoming skeptical.

: There already is a .Net Framework implementation (called mono) and C#
: compilers, etc. available for linux. And from the looks of it..they
: support most everything already. Maybe I'll give it a try one of these
: days and see what happens.

"they support most everything already. . . ."
^^^^
I can see where this is going. . . . .

: Though the rendering core is Managed DX to where there is currently no
: port available. But....those are confined to a single DLL which I
: could easily swap for a linux version equivalent.

"could easily swap for a linux version . . ."
^^^^^
Yeah, and "your check is in the mail."

I agree with Kevin here (but without the obscenities): You are
re-writing the same old CAD stuff in a new language which is
purportedly better. You are trying to gather feature requests from
potential customers, which is a good thing. But what's the real
advantage to the package? Better software technology? The customer
doesn't care.

And anyway, it's my opinion only, but basing your code on Microsoft
products is not necessarily going to product a stable, more robust
program. It *will* lock your customers into that platform at a time
when lots of people are trying to defect. This is not a good selling
point.

Of course, I am just a single data point. I am always interested in
hearing from others about whether they want CAD on Linux over Windoze,
and whether they would actually pay for it -- which is always the
sticking point with Linux stuff.

:>: I highly appreciate the input though, and will regardless check those
:>: things out just to see what they do and how. Can only learn.

Good attitude! Please post a URL to your product when it is ready!

Stuart

 

[Stuart Brorson]

: Stuart Brorson wrote:

:> Even today, MicroSim's integration of SPICE analysis with schematic
:> capture hasn't been beat.

: I have to disagree with this

So can SuperSpice do multi-layer PCB layout?

Stuart

 

[Stuart Brorson]

: I agree about the mass migration to Linux (the more the merrier), but we
: decided to take a multi-platform approach. Our software is written in Delphi
: 6.0/Kylix 2.0 for the frontend and C++ for stuff that is not OS specific.

If you are doing multi-platform development, Delphi/Kylix & C++ are much
more sensible language choices than C# and .nut.

Stuart

 

[Stephan Rose]

Engineers like to complain. Engineers *always* think they can do it
better. They wouldn't be engineers if that weren't the case. However,
the naive oes exclude time and money in their evaluations.


Oh dear...the sucker..

For....doing my job? mmm...ok.

All in no time with zero cost.

Actually how about a year and my salary as cost?

http://www.geda.seul.org/

If you think it doesn't have enough features, then why not take the
existing code base and add them?

As for PCB layout, there is already an existing program: PCB:

http://pcb.sourceforge.net/

PCB's GUI is primitive. If you feel like hacking, why not take PCB's
innards and build them into a GUI compatible with the 21st century?

IMHO, there are too many 1/2 completed free software projects out
there. Why not take one that already exists and extend it? (Note
that gEDA is already 100% functional, but could always use new
features. . . .)

[ . . . description of EDA project snipped . . . .]
Here's the list of in what order I plan on implementing what

1. Core engine and functions - Current stage
2. Schematic Component Editor - Current stage
3. Schematic Design
4. Pcb Component Editor
5. Pcb Design

So there it is...now how about some input?

Ok... here is is. You won't like it though. Look, every one of us has
these grandiose ideas about making some sort of great, super duper
perfect product. It will *never* happen. This is the real world.

Personally, I thing you and your company are completely wasting your
time. It *will* be time and money directly down the toilet. There is way
to much competition in general purpose cad for a new entry to make it.
You wont take sales away from the existing customers. The only way you
will get marketing share is by giving it away.

Our initial target won't be the high end customers that can fork out
10+ grand per license....it'll be the low end customers who can't
afford the high end stuff, and are frustrated on the low end side..oh
there's quite a market share on the low end market if you make it
affordable enough to where someone can afford to buy it w/out taking
too many risks.

You seem to have the daft and naive idea that people buy and use the
"best" products. They don't. Secondly, you product can't possible be the
"best" because everyone wants something *different*. Its impossible to
cover all bases, this is so, even if the software writer is an *expert*
in the actual *use* of such software

And when did I say I was writing the "best"? I'm simply trying to be
good, and trying to improve on things I personally saw that could use
improvement, and currently trying to gather on ideas that other people
think could use improvement and see if any of it can be included.

This means you have a fundamental issue with evaluating the business
sense of doing this product.


I think you just have not done enough market research on this. Its
flooded. Realistically, there is *zero* chance of a new same shit,
different day cad product making it today.

Even so..we've not found a single (affordable) product so far that has
not made us curse within the first 5 minutes and hit the uninstall
button a minute later.

I suspect you wont, as people generally only like positive comments,
they ignore the negative ones. People seeking advice are usually only
seeking an accomplice.

Actually I have no problems with your input. They are appreciated just
as much...

So, you *already* have decided on this idea of doing a product, that 30+
other companies already make, and yet haven't really seen what is
already out there and *why* they are being used.

I've seen the products we've tested. I've seen the product we use.
Hence why we arrived where we are right now, we've not been happy with
anything we've tested so far.

 

[Kevin Aylward]

For....doing my job? mmm...ok.

eell, this was before I read below to see that you were getting paid for
it. Now its skimming what you can off your employer by supporting
something of benefit to you, but realistically, of no real value to your
company.

Actually how about a year and my salary as cost?

No chance. It will take way longer than 1 man year. Trust me on this.
The SS gui is 100k lines of code, and no, its not bloated.

wrote:

[ . . . description of EDA project snipped . . . .]
Here's the list of in what order I plan on implementing what

1. Core engine and functions - Current stage
2. Schematic Component Editor - Current stage
3. Schematic Design
4. Pcb Component Editor
5. Pcb Design

So there it is...now how about some input?

Ok... here is is. You won't like it though. Look, every one of us has
these grandiose ideas about making some sort of great, super duper
perfect product. It will *never* happen. This is the real world.

Personally, I thing you and your company are completely wasting your
time. It *will* be time and money directly down the toilet. There is
way to much competition in general purpose cad for a new entry to
make it. You wont take sales away from the existing customers. The
only way you will get marketing share is by giving it away.

Our initial target won't be the high end customers that can fork out
10+ grand per license....it'll be the low end customers who can't
afford the high end stuff, and are frustrated on the low end side..oh
there's quite a market share on the low end market if you make it
affordable enough to where someone can afford to buy it w/out taking
too many risks.

But there are already too many companies in the low end market. They
*all* had the same idea. Why do you think that you can succeed where
these, supposedly fail?. Are you an personally an expert pcb layout and
schematic user in conjunction with your software abilities? Do you know,
personally, exactly *all* of the details that are required for a good
pcb product e.g gerber formats, routing algorithms etc... If not, no
chance. One has to make mistakes. Its the only way we get things
correct.

e.g. see http://www.anasoft.co.uk/replicators/index.html

Its the bit about random generation and selection.

And when did I say I was writing the "best"?

It was implied in the overall flow of your post.

I'm simply trying to be
good, and trying to improve on things I personally saw that could use
improvement, and currently trying to gather on ideas that other people
think could use improvement and see if any of it can be included.

This means that you don't *already* know what makes a good cad product,
so how to you think that you are going to overcome this learning curve?
What do you think all the other companies do? You reckon that they never
ask their customers what they would like?

Even so..we've not found a single (affordable) product so far that has
not made us curse within the first 5 minutes and hit the uninstall
button a minute later.

You get what you pay for, more or less. What I will tell you, is that
when I increased my product sale price from $50 to $200, I truly got
more sales. Think some more on this.

I've seen the products we've tested. I've seen the product we use.
Hence why we arrived where we are right now, we've not been happy with
anything we've tested so far.

What your implying is that the 100,000's of thousands, and I mean
100,000's of companies using schematic capture/pcb products are not
shipping product. The cost of cad software for any companies but the 2
man, work in bedroom job, is pretty much irrelevant. What's $5 grand? A
months salary? That's why products are priced at what they are. Its what
the market can stand. You simply aren't going to make money by selling
to the cheapskates. All you'll do is make the other less able companies
sell for even less, and these cheapskate companies, will get by with
those products.

I wrote SS for personal satisfaction. Its something I personally wanted
to do. 4 years of coding was a nice little hobby. However, for a
professional company to undertake, "yet another me too product", is
pretty much daft, imo.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

"That which is mostly observed, is that which replicates the most"
http://www.anasoft.co.uk/replicators/index.html

"quotes with no meaning, are meaningless" - Kevin Aylward.

 

[Kevin Aylward]

"could easily swap for a linux version . . ."
^^^^^
Yeah, and "your check is in the mail."

I agree with Kevin here (but without the obscenities):

And just what alledged obscenities are you refering to?

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

"That which is mostly observed, is that which replicates the most"
http://www.anasoft.co.uk/replicators/index.html

"quotes with no meaning, are meaningless" - Kevin Aylward.

 

[Kevin Aylward]

So can SuperSpice do multi-layer PCB layout?

As I explained, there is no one tool, that can can all, well. That why I
ignore pcb layout.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

"That which is mostly observed, is that which replicates the most"
http://www.anasoft.co.uk/replicators/index.html

"quotes with no meaning, are meaningless" - Kevin Aylward.

 

[Stephan Rose]

: On Wed, 04 Feb 2004 02:23:01 -0000, Stuart Brorson <[email protected]>
: wrote:
:>OK, fine. I assume that you are targeting the low end, because there
:>are already enough tools at the high end (i.e. Viewdraw, Allegro,
:>Mentor's stuff, etc.). Also, the big players (e.g. Cadence) are
:>ababdoning the low end, leaving market opportunity.

: That is going to be our initial target. Going to expand towards the
: high end as time goes on. But low end is our first and initial
: priority.

I tend to agree with Kevin's point (elsewhere on this thread) that the
CAD market is flooded, particularly at the low end. YOu may have
trouble making money on this. That being said, I don't know of many
end-2-end suites which do schematic capture, SPICE simulation, *and*
decent, multi-layer layout. If I am missing one, please let me know.
For example, does Kevin's SPICE prog do layout?

An aside about your project: All file should be easily parsable
ASCII. Hiding things in binary files, or -- worse -- in "databases"
(a la Protel) is anathema.

: Hence why I'm using the language and envrioment I am using.

: It's not quite as fast as C++, but in exchange it offers extreme
: stability. Invalid memory accesses, memory leaks, etc. (and crashes
: caused by that) are virtually impossible. No pointer mess either. It's
: beautiful

Sounds like Java. . . .

: Actually C# and the .Net Framework are designed by microsoft to be
: cross-platform and easily portable. When compiled, the code generated
: actually is not executable code, but intermediate byte code instead.
: Hence, my software would run on any operating system and hardware
: supporting the .net framework. The actual executable code is created
: at run-time on the target platform the executable is run on.

Sounds like Java. . . . .

So what's to prevent it from being sloooooooooow, like Java? And
what's to prevent Microsoft lock-in, as C# & .nut morph ever so
slightly over time to keep everybody off balance unless they stick
with the latest version of Windoze? This is what has happened with
InternetExploiter and the various versions & file formats of
MSOffice. And this is also true of C# -- it's really just Microsoft's
attempt to derail Java by providing an equivalent language which only
Microsoft controls.

As a potential customer, I am already becoming skeptical.

Because the C# language standard is explicity designed to be
cross-platform. As far as office and such is concerned, well doesn't
virtually every program out there have that issue? file formats keep
evolving so you keep needing the latest version of the software? That
applies to MS Office and virtually everything else out there....

As far as execution speed is concerned....while the generated
executable isn't actual executable code, at run-time, the first time
any program is run..the executable code for it is created...and from
thereon, it behaves just like any other standard C/C++ equivalent
program, including execution speed. The only reason a program written
in this envrioment might be slightly slower than C/C++ equivalent is
all the added security (ie access to arrays are validated to ensure
they're not out of bounds, and so on). This means a very stable
program, at a slight performance cost though. Such is the world of
coding..nothing is free.

Though having written in this envrioment for the past 2 years....I've
never had performance problems.

: There already is a .Net Framework implementation (called mono) and C#
: compilers, etc. available for linux. And from the looks of it..they
: support most everything already. Maybe I'll give it a try one of these
: days and see what happens.

"they support most everything already. . . ."
^^^^
I can see where this is going. . . . .

: Though the rendering core is Managed DX to where there is currently no
: port available. But....those are confined to a single DLL which I
: could easily swap for a linux version equivalent.

"could easily swap for a linux version . . ."
^^^^^
Yeah, and "your check is in the mail."

I agree with Kevin here (but without the obscenities): You are
re-writing the same old CAD stuff in a new language which is
purportedly better. You are trying to gather feature requests from
potential customers, which is a good thing. But what's the real
advantage to the package? Better software technology? The customer
doesn't care.

Here's some things I can think of on the top of my head that I've not
seen yet... (I'm sure a lot of this is present in very high end
applications..we've just not seen it on the low end so far).

Well one advantage I intend to have is one completely integrated
package. ONE application....But, flexible in its features via plugin
technology. So the same application could be shipped with different
features and very easily upgraded w/out reinstalls or requiring
different apps. Just add the appropriate plugin files.

A system where....if you update the schematic...hit a button...netlist
for the pcb will automatically update as well. Ability to have
multiple boards under 1 project, etc.

Also...being the 3D nut I am...down the line....ability to define your
components in 3d, put the boards together (if say you have a design
that has multiple boards attached in some weird configuration to each
other) and make sure things fit, components don't interfere, etc. This
would be at very later stages pretty far down the line though.

And anyway, it's my opinion only, but basing your code on Microsoft
products is not necessarily going to product a stable, more robust
program. It *will* lock your customers into that platform at a time
when lots of people are trying to defect. This is not a good selling
point.

Actually I'm not locked....C# itself is just a language and therefore
has no operating system/hardware bounds. Also, while created primarily
by microsoft, it is a standardized language just like C/C++. You can
read about it here if you're interested. (PDF at the bottom) And I
suppose you can view the .Net Framework as C#'s version of the STDLIB
of C/C++.

http://www.ecma-international.org/publications/standards/Ecma-334.htm

The .Net framework is nearly fully supported under linux with the only
things missing are some rare obscure things that nobody uses
anyway....

To change the rendering core to a linux equivalent....could do that in
a week as the application code itself makes no api specific calls, so
all api specific code is confined to a single location only..I think
totalling 20 lines of code or so

Of course, I am just a single data point. I am always interested in
hearing from others about whether they want CAD on Linux over Windoze,
and whether they would actually pay for it -- which is always the
sticking point with Linux stuff.

:>: I highly appreciate the input though, and will regardless check those
:>: things out just to see what they do and how. Can only learn.

Good attitude! Please post a URL to your product when it is ready!

Thanks I'm going to have some initial screenshots of the SCM
Component designer up later this month...I'll post some links to that.

 

[Stephan Rose]

No chance. It will take way longer than 1 man year. Trust me on this.
The SS gui is 100k lines of code, and no, its not bloated.

Longer for the end goal, yes. But to have a sellable product to start
out with, 1 year is quite possible.

As far as the lines of code are concerned, great for you...I'm not
gonna get into a code line contest here.....

But there are already too many companies in the low end market. They
*all* had the same idea. Why do you think that you can succeed where
these, supposedly fail?. Are you an personally an expert pcb layout and
schematic user in conjunction with your software abilities? Do you know,
personally, exactly *all* of the details that are required for a good
pcb product e.g gerber formats, routing algorithms etc... If not, no
chance. One has to make mistakes. Its the only way we get things
correct.

If everyone though "jeez..nobody else was able to do this..so I can't
either." you know where we would be? we'd still be sitting in caves,
banging on rocks hoping we'll get a spark for a fire....

No I don't know *ALL* the details. I probably don't even know *half*
the details. But I'm willing to learn, willing to listen, willing to
create a product, willing to bring it out..willing to listen to
feedback, implement the feedback..and improve.

It was implied in the overall flow of your post.

If that's your interpretation of it...

This means that you don't *already* know what makes a good cad product,
so how to you think that you are going to overcome this learning curve?
What do you think all the other companies do? You reckon that they never
ask their customers what they would like?

The same way I overcome every other learning curve. I climb it...

You get what you pay for, more or less. What I will tell you, is that
when I increased my product sale price from $50 to $200, I truly got
more sales. Think some more on this.

And I am pushing to charge as much as possible..while still trying to
maintain a little edge price and feature wise.

I wrote SS for personal satisfaction. Its something I personally wanted
to do. 4 years of coding was a nice little hobby. However, for a
professional company to undertake, "yet another me too product", is
pretty much daft, imo.

And that's your opinion, you're very much entitled to it.

 

[Charles Edmondson]

Because the C# language standard is explicity designed to be
cross-platform. As far as office and such is concerned, well doesn't
virtually every program out there have that issue? file formats keep
evolving so you keep needing the latest version of the software? That
applies to MS Office and virtually everything else out there....

As far as execution speed is concerned....while the generated
executable isn't actual executable code, at run-time, the first time
any program is run..the executable code for it is created...and from
thereon, it behaves just like any other standard C/C++ equivalent
program, including execution speed. The only reason a program written
in this envrioment might be slightly slower than C/C++ equivalent is
all the added security (ie access to arrays are validated to ensure
they're not out of bounds, and so on). This means a very stable
program, at a slight performance cost though. Such is the world of
coding..nothing is free.

Though having written in this envrioment for the past 2 years....I've
never had performance problems.




Here's some things I can think of on the top of my head that I've not
seen yet... (I'm sure a lot of this is present in very high end
applications..we've just not seen it on the low end so far).

Well one advantage I intend to have is one completely integrated
package. ONE application....But, flexible in its features via plugin
technology. So the same application could be shipped with different
features and very easily upgraded w/out reinstalls or requiring
different apps. Just add the appropriate plugin files.

A system where....if you update the schematic...hit a button...netlist
for the pcb will automatically update as well. Ability to have
multiple boards under 1 project, etc.

Also...being the 3D nut I am...down the line....ability to define your
components in 3d, put the boards together (if say you have a design
that has multiple boards attached in some weird configuration to each
other) and make sure things fit, components don't interfere, etc. This
would be at very later stages pretty far down the line though.




Actually I'm not locked....C# itself is just a language and therefore
has no operating system/hardware bounds. Also, while created primarily
by microsoft, it is a standardized language just like C/C++. You can
read about it here if you're interested. (PDF at the bottom) And I
suppose you can view the .Net Framework as C#'s version of the STDLIB
of C/C++.

http://www.ecma-international.org/publications/standards/Ecma-334.htm

The .Net framework is nearly fully supported under linux with the only
things missing are some rare obscure things that nobody uses
anyway....

To change the rendering core to a linux equivalent....could do that in
a week as the application code itself makes no api specific calls, so
all api specific code is confined to a single location only..I think
totalling 20 lines of code or so




Thanks I'm going to have some initial screenshots of the SCM
Component designer up later this month...I'll post some links to that.

Stephan,
Take a look at Microwave Office from AWR. You will see a lot of what
you are thinking of doing, except they initially targeted the RF market
for their products. Real Cool Stuff!

Charlie
Edmondson Engineering
Unique Solutions to Unusual Problems

 

[Jim Thompson]

Stephan Rose wrote: [snip]

Thanks I'm going to have some initial screenshots of the SCM
Component designer up later this month...I'll post some links to that.

Stephan,
Take a look at Microwave Office from AWR. You will see a lot of what
you are thinking of doing, except they initially targeted the RF market
for their products. Real Cool Stuff!

Charlie
Edmondson Engineering
Unique Solutions to Unusual Problems

Speaking of which... how's Mohi enjoying his position at AWR?

...Jim Thompson