Text vs binary

[martin_gnu]

Franz Reiter ask us about this...

Masterful document for all !!!! (page 15 and page 17)
http://www.ncsa.uiuc.edu/NARA/XML_and_Binary.pdf

Interesting discussion:
http://www.thescripts.com/forum/thread86383.html
tutorial about assci and binary files:
http://www.cs.umd.edu/class/spring2003/cmsc311/Notes/BitOp/asciiBin.html

Regards.

[Lars Grobe]

> Interesting discussion:
> http://www.thescripts.com/forum/thread86383.html

Actually I found most of it quite strange. Saving preferences in bytecode because utf-16 would have portability problems? What are the guys talking about? The whole world saves more or less everything in utf-8 and the guy cannot get his preferences file into it? And what does it mean for CAD? You can encode whatever you want both binary and ascii, and I would certainly use binary only as a performance workaround. I have used grep, sed & co on CAD files and the binaries were uncomfortable enough to make me even think about it.

Sorry I have to insist, what are the aims, target use of a new format?

Lars.

[martin_gnu]

> Actually I found most of it quite strange. Saving preferences in bytecode because utf-16 would have portability problems? What are the guys talking about? The whole world saves more or less everything in utf-8 and the guy cannot get his preferences file into it? And what does it mean for CAD? You can encode whatever you want both binary and ascii, and I would certainly use binary only as a performance workaround. I have used grep, sed & co on CAD files and the binaries were uncomfortable enough to make me even think about it.
>
> Sorry I have to insist, what are the aims, target use of a new format?
>

"An agreement between all CAD developers"
I would like people could defend a only file especification. Nowadays
all free CAD developers use a diferent CAD format to work:
Qcad (DXF)
gcad (gcad native)
BRL-CAD (.g database)
...

I like text also (pattern matching in PERL) but people would like open
a CAD file in the best time.

the most important link is:
http://www.ncsa.uiuc.edu/NARA/XML_and_Binary.pdf

They talk about binary files in engineering. We always work with
doubles-Integers and if we use text files we always do conversions. In
engineering we have more numbers than text, and therefore like Franz
Said could be interesting study that posibilibity also.

BRL-CAD (U.S. Army) uses a binary file. They have especification and
they are many people working in that project.

Regards.

[Lars Grobe]

> "An agreement between all CAD developers"
> I would like people could defend a only file especification. Nowadays
> all free CAD developers use a diferent CAD format to work:
> Qcad (DXF)
> gcad (gcad native)
> BRL-CAD (.g database)

But this is related to the information contained. QCad is 2d, gcad is geometry-centric afaik, brlcad is made for solid modeling and hold lots of data. If you want ONE format, representing everything from 2d sketches to part numbers, material properties etc, things won't be too efficient. That is why step is modular btw. And just forget if it comes to paramertic systems where you do not want to store geometry but e.g. number of raisers, material/construction and width of a stair so that this can be generated by the app. So if you want to have ONE native format for all CAD this will be a monster, especially if monolythic. If modular, it will be something like reinventing STEP I guess.

> BRL-CAD (U.S. Army) uses a binary file. They have especification and
> they are many people working in that project.

For moving data between platforms there is a text format, the standard database is binary.

I have doubts about the idea of a general native format. If you are thinking about exchange formats fine. If it is for sharing of information, fine. But native without specified use is difficult. What is CAD?

Lars.

["Martín RV (OPENGeoMap)"]

> So if you want to have ONE native format for all CAD this will be a monster, especially if monolythic. If modular, it will be something like reinventing STEP I guess.
>
>

Yes. i didn´t know all differents CADs... Perhaps you are true, but,
don´t you think that we have a problem in the free cad world?.

I thought that we only have Autocad-Microstation, but we have Catia also...
How many different kind CAD vendors we have?

You told me about Qcad, but i think that all CADs can be in 3D. I don´t
think that we need a file format for 2D and other for 3D.

We have the different CADs here:
http://ftp.brlcad.org/Industry_Diagram.png

But 1 file format in not crazy. For example:
a).- We have sphere primitive in a file.
b).-Our CAD don´t have this primitive. Well, the library to import the
file could break the sphere primitive in Nurvs or other compatible
complex primitive in that case. I think that is important the file
format and a library to use that file format. All people could use the
file format in the same way because all use the same library tu use it.

Is this very crazy?

Regards.

[Franz Reiter]

Hallo,

concerning:

> > Sorry I have to insist, what are the aims, target use of a new format?

why not define the goals?
- geometric constructions in 2D (draftings)and 3D
- geometric constructions in 3D; for engineering, architucture, GIS
- viewing cad-data, gis-data, vitual reality-data (mesh+images)
- measuring functions
- functions for documentations
- ... ???

ad:

>> I would certainly use binary only as a performance workaround

>> In engineering we have more numbers than text ..
at the moment we are testing a LandXML-importprogram. We can load (display)
200000 triangles in ~ 0.2 sec (from ascii !) translate it forward and back
into ascii for permanent save.
It seems that we can prove that loading from ascii does not -
- slow down (too much) and does not
- make bigger storagesize than binary.

>> So if you want to have ONE native format for all CAD this will be a
monster, especially if monolythic. If modular, it will be something like
reinventing STEP I guess.

step is complicated, for non-members nearly secret, but the best approach at
the moment.
So, why not reinvent/improve step ?
And put the definitions into something like a wiki .. ??

Franz

[Lars O. Grobe]

> why not define the goals?
> - geometric constructions in 2D (draftings)and 3D
> - geometric constructions in 3D; for engineering, architucture, GIS
> - viewing cad-data, gis-data, vitual reality-data (mesh+images)
> - measuring functions
> - functions for documentations

Good start... so the focus is on geometry (2d&3d) with related functions
for documentation and analysis. Now it comes to questions such as brep,
csg etc...

I doubt that these goals can be combined with vr, but a good x3d-export
would be enough. If the format is to include vr, all the stuff that took
years to get approved for x3d (and is accepted there) had to be
reinvented, and things like behaviour and network functionality may be
just too much for parsing a CAD file. Also vr-formats usually are
graph-organized. And it takes a long time to establish a standard in
this market, especially if there is an ISO-backed standard.

> step is complicated, for non-members nearly secret, but the best approach at
> the moment.
> So, why not reinvent/improve step ?

My problem with reinventing is that I doubt that it will be easier then
pushing forward existing free step implementations instead. And this
would, at the same time, guarantee interoperability with existing
software and with commercial applications.

I once hoped that one common format would become possible. I thougth
about using x3d with some additions, basically a new profile for CAD.
But e.g. the graph structure seams to be not very efficient for CAD
editing. Than I would need more geometry and less behaviour / network in
CAD. And finally I came to a point that I would be by far more
interested in a kind of scripting environment that could use modules for
e.g. stairs, walls, windows, and save only the calls to the scripts
instead the resulting geometry. I am in architecture, and pure geometry
is rather dumb for architecture. I need to define walls windows etc, and
than output 3d-models, 2d-drawings, material lists etc from this
information. Call it parametric and application-specific. These
questions seam to be more important to me now then how to save it on a
hard disk. Maybe I would prefer never to directly save it on a hard
disk, but store my objects in a ldap-tree (with caching of course) and
allow sharing, access control and replication.

CU Lars.

[martin_gnu]

> Good start... so the focus is on geometry (2d&3d) with related functions
> for documentation and analysis. Now it comes to questions such as brep,
> csg etc...

This is a old question i think
http://wheger.tripod.com/step/

>
> My problem with reinventing is that I doubt that it will be easier then
> pushing forward existing free step implementations instead. And this
> would, at the same time, guarantee interoperability with existing
> software and with commercial applications.

I think it´s posible and necessary reinvent the wheel. we only need
defend a only CAD format. We are many people...
We need:
- 1 format
- libraries to work with this format (c, PERL, JAVA,python...)
- people working in conversions routines between this format and X3d,
STEP, DWG, DGN, DXF,...

It´s a long way but i think that it´s the best way.

> CAD. And finally I came to a point that I would be by far more
> interested in a kind of scripting environment that could use modules for
> e.g. stairs, walls, windows, and save only the calls to the scripts
> instead the resulting geometry. I am in architecture, and pure geometry
> is rather dumb for architecture. I need to define walls windows etc, and
> than output 3d-models, 2d-drawings, material lists etc from this
> information. Call it parametric and application-specific. These

This is the reason for which open_cad_format began. We think that the
only complete format is DWG, but we depend on the great taliban of the
engineering (Autodesk). If you makes plugins in Autocad you can
understand that DWG allows you to create compound primitives . You can
create the primitive one with the "window" or "walls" names like
your you say.
There are a lot of CAD plugins verdors of Autocad working with this
compound primitives, but Autodesk have the best plugins like Autodesk
Civil for example.

Autodesk is the great TALIBAN for engineers...
http://usa.autodesk.com/adsk/servlet/index?id=331041&siteID=123112

I believe that we don´t have to compare STEP with DWG, because STEP is
a exchange format. It´s not useful for a real work like X3D.

I am talking by mail with Richard Stallman and i hope good news...
Regards.

[Lars O. Grobe]

>> My problem with reinventing is that I doubt that it will be easier then
>> pushing forward existing free step implementations instead. And this
>> would, at the same time, guarantee interoperability with existing
>> software and with commercial applications.
>

> I think it愀 posible and necessary reinvent the wheel. we only need

> defend a only CAD format. We are many people...

There are two questions regarding this:

1) Is it necessary? What are the advantages / disadvantages over using
step? The only real argument for a new format here was simplicity, and I
wonder if a new format will still be simple if it covers everything from
2d sketches to parametric modeling (application specific).

2 Is it possible? This is a technical question as well as one about
timeframe, ressources, manageability.

> This is the reason for which open_cad_format began. We think that the
> only complete format is DWG, but we depend on the great taliban of the
> engineering (Autodesk). If you makes plugins in Autocad you can
> understand that DWG allows you to create compound primitives . You can
> create the primitive one with the "window" or "walls" names like
> your you say.

I use DWG only for 2d line drawings, have no idea about the concept of a
"compound format". The formats I used where always application specific.
If you want to open a drawing containing a parametric model of a stair
in an Archicad-file e.g., you need the procedures necessary to create
the stairs geometry from parameters. In Archicad, this is solved either
by bundled modules or by scripts. I guess the concepts could work with
open cad formats if the scripts' language was not propietary but open,
such as java, python etc. Than you could save a routine needed to
interpret parameters together with these parameters and regular
geometry. I do not know if step can do something like this.

> I believe that we don愒 have to compare STEP with DWG, because STEP is
> a exchange format. It愀 not useful for a real work like X3D.

I guess a lot of real work is done in x3d. Not CAD drawing work.

CU Lars.

[martin_gnu]

>
>There are two questions regarding this:
>
>1) Is it necessary?
>

>2 Is it possible?
>
What are you ideas then?.
Use STEP, X3D, DXF, DWG...

>
>I use DWG only for 2d line drawings, have no idea about the concept of a
>"compound format".
>

For example.
I use a plugin called TCP-IP (for surveying) and they created a entity
in DWG called punto3D:
punto3D= point + text +text+ block
In Autolisp or Visual Basic you can find that entity... but we want
defend PERL, python like you say.

>The formats I used where always application specific.
>

Okay. There are different opinions. Some people prefer Autocad+plugins
and other people prefer specific CADs. I prefer Autocad+plugins

It´s Archicad or Autocad+plugins better for Architecture ???

>If you want to open a drawing containing a parametric model of a stair
>in an Archicad-file e.g., you need the procedures necessary to create
>the stairs geometry from parameters. In Archicad, this is solved either
>by bundled modules or by scripts. I guess the concepts could work with
>open cad formats if the scripts' language was not propietary but open,
>such as java, python etc. Than you could save a routine needed to
>interpret parameters together with these parameters and regular
>geometry. I do not know if step can do something like this.
>

I don´t know anything about Architecture, but is it posible solve the
problems with compound entities?
I believe that in the future the FREE/OPEN CADs will have scripting
posibilities like Microstation o Autocad.
http://www.swig.org/
SWIG is a software development tool that connects programs written in
C and C++ with a variety of high-level programming languages such as
Perl, PHP, Python, Tcl, Ruby and PHP. The list of supported languages
also includes non-scripting languages such as C#, Common Lisp (CLISP,
Allegro CL, CFFI, UFFI), Java, Modula-3 and OCAML.

Autodesk has Architecture software:
http://usa.autodesk.com/adsk/servlet/item?siteID=123112&id=2773513

I don´t know nothing about Archicad or Architecture , but for many
people Autodesk solve all problems in DWG format.

We have 4 billions of DWG files nowadays (acording to open alliance).
Regards.

[Franz Reiter]

Hallo;

I forgot 2 important features, so the completed list is:

- geometric constructions in 2D (draftings)and 3D
- geometric constructions in 3D; for engineering, architucture, GIS
- viewing cad-data, gis-data, vitual reality-data (mesh+images)

- scripting
- NC-programming

- measuring functions
- functions for documentations

With scripting and parametric one could describe all kinds of geometry (eg
windows, stairs for architectural applications). The connection to
material-list and part-lists needs more discussions..

>My problem with reinventing is that I doubt that it will be easier then
>pushing forward existing free step implementations instead.

i do not know step good enough to know if is possible to hold a program for a
stair in a step-model, but i think no ...

Ad VR:
vr shall not be the primary focus; but the basic functions (applying textures,
moving around) should be possible without too much effort ..
If one needs more: exportFunction to x3d...

Franz

[martin_gnu]

>
>
>Hallo;
>
>I forgot 2 important features, so the completed list is:
>- geometric constructions in 2D (draftings)and 3D
>- geometric constructions in 3D; for engineering, architucture, GIS
>- viewing cad-data, gis-data, vitual reality-data (mesh+images)
>- scripting
>- NC-programming
>- measuring functions
>- functions for documentations
>

I put this in the home page...
http://groups.google.com/group/open_cad_format

I update the free libraries list with SWIG:
http://groups.google.com/group/open_cad_format/web/free-libraries

>
>With scripting and parametric one could describe all kinds of geometry (eg
>windows, stairs for architectural applications). The connection to
>material-list and part-lists needs more discussions..
>

Perhaps in databases (sqlite, Mysql, postgree,XML...)
Many GIS aplications over Autocad or Microstation use databases
connections...
it´s a good choice SWIG to scripting .
SWIG was created by Los alamos:
http://en.wikipedia.org/wiki/Los_Alamos_National_Laboratory
It´s a OSgeo recomendation...
http://wiki.osgeo.org/index.php/Programming_Languages_Project

>
>
>Ad VR:
>vr shall not be the primary focus; but the basic functions (applying textures,
>moving around) should be possible without too much effort ..
>If one needs more: exportFunction to x3d...
>
>
>

I totally agree. We have blender and K-3D to that task,...
A CAD is totally different to a 3D modeling sofware...
K-3D is a 3d modeling software created by Lucas Digital Arts (Harry
potter, Star Wars,...)

I believe that we have 3 kind of software:
- CAD (We are here talking about this...)
- 3d modeling software (vr, Games, visualization, films, special
effects, simulations,...)
- 2d design like Freehand, illustrator, inskape, corel draw,...

Regards.

[Lars O. Grobe]

Hi!

> What are you ideas then?.
> Use STEP, X3D, DXF, DWG...

As I wrote, I would prefer using step, which is an open cad format, and
put all efforts in developing an open source framework to get it into
all open source projects dealing with design relevant data. There are
two questions before I can claim that this would be useful. First is,
does step provide us with enough functionality, e.g. for parametric and
scripting stuff, to be used in advanced engineering and design
applications? Second is, how easy / difficult will this task be, and
will it be easier or more difficult than developing a new standard
covering all the functionality and getting it accepted and approved (ISO?)?

>> I use DWG only for 2d line drawings, have no idea about the concept of a
>> "compound format".
>>
> For example.
> I use a plugin called TCP-IP (for surveying) and they created a entity
> in DWG called punto3D:
> punto3D= point + text +text+ block
> In Autolisp or Visual Basic you can find that entity... but we want
> defend PERL, python like you say.

So you mean the ability to have entities that are structures made of
other entities? I guess this is a rather basic feature for a CAD format,
so we should not worry to much.

>> The formats I used where always application specific.
>>
> Okay. There are different opinions. Some people prefer Autocad+plugins
> and other people prefer specific CADs. I prefer Autocad+plugins

For architecture, there used to be Architectural Desktop wich was
something on top of a regular Autocad installation. I do not know if
current releases still work like that or are more or less independent
applications. Actually, people seam to move towards Revit if using
Autodesk products. Most people I know in architecture use Autocad only
in 2d as a drawing tool and use other parametric software when working
in 3d. Just imagine the case that the thickness of your exterior walls
changes and you had to change all doors, windows etc in classical
Autocad - in 2d this can still be done somehow, in 3d you would start
from zero.

> I believe that in the future the FREE/OPEN CADs will have scripting
> posibilities like Microstation o Autocad.

Most CADs and modelers already have. But the scripts usually do not
survive import/export or even don't exist in the native format. On
example from modeling is Blender - you have scripts for everything. But
these scripts generate geometry, so not the call to the script is saved
but its output. This means, for my stairs-example, that you can create a
stair with a script, but not change it later by changing parameters.

BTW. The idea of the big and for-everything format completely
contradicts Eric Wilhelm's vectorsection work (compact native formats,
big hub format). We should be aware of this, he is a clever guy and has
thought and discussed this topic for years.

CU Lars.

["Martín RV (OPENGeoMap)"]

Hi:

As I wrote, I would prefer using step, which is an open cad format, and 
put all efforts in developing an open source framework to get it into 
all open source projects dealing with design relevant data. There are 
two questions before I can claim that this would be useful. First is, 
does step provide us with enough functionality, e.g. for parametric and 
scripting stuff, to be used in advanced engineering and design 
applications? Second is, how easy / difficult will this task be, and 
will it be easier or more difficult than developing a new standard 
covering all the functionality and getting it accepted and approved (ISO?)?
  

Perhaps you are true. Perhaps i was thinking about the posibility of create THE CAD, but there are a lot of free CADs. But i  think that we make efforts to build a great CAD for all. Nowadays we don´t have a profesional CAD.

For architecture, there used to be Architectural Desktop wich was 
something on top of a regular Autocad installation. I do not know if 
current releases still work like that or are more or less independent 
applications. Actually, people seam to move towards Revit if using 
Autodesk products. Most people I know in architecture use Autocad only 
in 2d as a drawing tool and use other parametric software when working 
in 3d. Just imagine the case that the thickness of your exterior walls 
changes and you had to change all doors, windows etc in classical 
Autocad - in 2d this can still be done somehow, in 3d you would start 
from zero.

  

This is a good reason. justifications to be efficient working in a CAD.

Most CADs and modelers already have. But the scripts usually do not 
survive import/export or even don't exist in the native format. On 
example from modeling is Blender - you have scripts for everything. But 
these scripts generate geometry, so not the call to the script is saved 
but its output. This means, for my stairs-example, that you can create a 
stair with a script, but not change it later by changing parameters.
  

Yes scripts are internal to programs, this is the reason for defend THE CAD. Perhaps a dream. I don´t know but  if people like photography join to the GIMP team.

BTW. The idea of the big and for-everything format completely 
contradicts Eric Wilhelm's vectorsection work (compact native formats, 
big hub format). We should be aware of this, he is a clever guy and has 
thought and discussed this topic for years.
  

Yes you are true. But Franz Reiter is a clever guy also, hehe. He is building a great CAD and working. We only want find the best choice between all. I believe that Eric Wilhelm opinions and Franz Reiter opinion are totally compatible. People of Archimedes is doing a great job in JAVA and we must think in them also. I believe this is a open forum, all can edit pages, all can give opinions,....
Don´t you think that free CAD situation is complicated nowadays?


Regards.

[Lars O. Grobe]

Hi!

> Perhaps i was thinking about the posibility of
> create THE CAD, but there are a lot of free CADs. But i think that we
> make efforts to build a great CAD for all. Nowadays we don´t have a
> profesional CAD.

It makes sense to work on a productive open source cad for design (I
think that is what you mean with "the CAD"). I would have liked this
open source CAD to use step as its native format, as this is an open
standard, supported by both commercial and free applications and afaik
an approved ISO standard.

If we talk about a new CAD (not a file format), we can talk about
concepts from both useage and technical points of view. I think this is
really important. I support it.

> This is a good reason. justifications to be efficient working in a CAD.

You know, the only reason people started to use CAD was efficiency ;-)
That is why having an useage scenario is important.

> Yes scripts are internal to programs, this is the reason for defend THE
> CAD. Perhaps a dream.

Combining data and procedures is not that new. Think of what happened to
html - all the script elements that are delivered with the content,
cross-platform, browser-independent. Just has to be done. It is not so
much about file formats, e.g. ecmascript was implanted into an existing
(html), too.

> Yes you are true. But Franz Reiter is a clever guy also, hehe. He is
> building a great CAD and working.

Yes, but I still wonder how to interprete that. Because actually, he
followed the "slim native" concept in his own way afaik. His
modifications to step to improve performance and simplify parsing as
much as possible and necessary for his app is just this kind of
application-specific optimization leading to a slim and unique format,
which can still be connected to an "uberformat" as Eric called it, the
hub. I am not sure, but that is what it appears to me.

> Don´t you think that free CAD situation is complicated nowadays?

Those who pay for free cad (in means of developer ressources etc) have
their free CADs and modelers, and most of them are very specialized and
often close to scientific / acedemic world. General engineering / design
companies did not put much into the development of such a CAD so far,
even not to talk about engineers, so nothing for them so far. E.g.
brlcad is increadibly advanced, but not useable for many of us. One app
that has made steps forward is Freecad, a nice start for a solid
parametric CAD. But it is unfinished. I am using propietary CAD as I do
not have an alternative yet.

I think one key point of creating a real open source general CAD is
thinking about development structures and funding.

CU Lars.

[Lars O. Grobe]

I once wrote down some ideas about developing a CAD environment. It was
ment for architects and civil engineers, but applies more or less to the
design process in general:

http://www.free-architecture.org/?q=node/12

["Martín RV (OPENGeoMap)"]

Hi:

I put your freearchitecture page here. It´s could be interesting for all:
http://groups.google.com/group/open_cad_format/web/free-cad-experiences

It makes sense to work on a productive open source cad for design (I 
think that is what you mean with "the CAD"). I would have liked this 
open source CAD to use step as its native format, as this is an open 
standard, supported by both commercial and free applications and afaik 
an approved ISO standard.
  

STEP is the best and complete format for all. I believe it then.
the last days i put a document here from National Center for Supercomputing Applications at the University of Illinois:
 (page 15 and page 17 are very interesting for us)
http://www.ncsa.uiuc.edu/NARA/XML_and_Binary.pdf

there are lot of miths about XML,plain text, binary...
I believe it´s important read this document.
In my experiences in C/GTK/GSL is a slow process the double to assci conversions. assci to double is really fast like Franz said the other day.
DWG format is encrypted and for that  it´s not so fast like DGN. If DWG wasn´t encryted you can open a big file and open-write in real time.

Combining data and procedures is not that new. Think of what happened to 
html - all the script elements that are delivered with the content, 
cross-platform, browser-independent. Just has to be done. It is not so 
much about file formats, e.g. ecmascript was implanted into an existing 
(html), too.
  

Myths. Only there are 3 html/javascript interpreters:
. Internet explorer
, Gecko (firefox, mozilla,...)
. webkit (safari, Khtml,...)

It´s imposible for us build a CAD interpreter from scrach. It´s a very difficult task.
We would need a Larry Wall... hehe

It´s more easy like Blender, gimp, K-3d, Autocad, microstation,... use external libraries to build scrips.
www.swig.org/
In swig you  can make plugins for your applications in the language that you want (c#,JAVA, python, PERL,...).

Yes, but I still wonder how to interprete that. Because actually, he 
followed the "slim native" concept in his own way afaik. His 
modifications to step to improve performance and simplify parsing as 
much as possible and necessary for his app is just this kind of 
application-specific optimization leading to a slim and unique format, 
which can still be connected to an "uberformat" as Eric called it, the 
hub. I am not sure, but that is what it appears to me.

  

The work of Eric is great. In the tree graph of vectorsection  i saw:
.blend is a binary/native  file of Blender. and k-3d binary/native???. K-3d is the software of lucas digital arts.
.dwg is a binary/native encrypted  file of Autocad.
.dgn is a binary/native file of Microstation.
.iges old file...
.step, assci file that it´s a good and complete CAD format.
Studying this files. Why bentley, autodesk, blender, k-3d don´t use open standards to work? perhaps because you open/write the file in almost real time?
why people uses png and jpeg  like open standars to photos and we don´t think about XML/assci to photos??
someone would want edit a photo in the notepad???
I don´t want begin a war between assci/xml/binary but it´s important think about this things. I don´t say that i don´t like STEP. If all people here defend STEP i would defend also.

parametric CAD. But it is unfinished. I am using propietary CAD as I do 
not have an alternative yet.
  

Like me. All people inside this list have the same problem. We need find solution to that...


Regards.

[Lars O. Grobe]

Hi!

> I put your freearchitecture page here. It愀 could be interesting for all:
> http://groups.google.com/group/open_cad_format/web/free-cad-experiences

Thanks, fine, it is a bit outdated, but I fear it is still more or less
valid in some points.

>> Combining data and procedures is not that new. Think of what happened to

>> html - all the script elements that are delivered with the content[...]

> Myths. Only there are 3 html/javascript interpreters:
> . Internet explorer
> , Gecko (firefox, mozilla,...)
> . webkit (safari, Khtml,...)
>

> It愀 imposible for us build a CAD interpreter from scrach.
[...]
> It愀 more easy like Blender, gimp, K-3d, Autocad, microstation,... use

> external libraries to build scrips.

> www.*swig*.org/

I do not want to propose inventing a new script language. I wanted to
point out that, if the file format is to be cross-application, and if
the data in these files depends on plug-ins, the plug-ins have to be
cross-application, too. Now there are two solutions for this (again see
html-browsers) - putting the routines into plug-ins that are
cross-application, or putting the routines into the file, together with
the data (as in ecmascript). I prefer the second, as actually usually
you have just that one plugin that is needed for opening the file not
installed on my computer ;-) BTW, you can already do such things, just
think a brlcad-file not being the "database" but the commands generating
the geometry.

> *.iges* old file...

Oh I really like it, allows nice data exchange between many modellers...

> why people uses png and jpeg like open standars to photos and we don愒

> think about XML/assci to photos??

There are some reasons for/against xml, far beyond text/bin (which is
only an encoding when writing to disk). Validation and filtering is far
less important for pictures, they are not a very structured set of data.
In CAD, having a xml encoding may be nice for (my thoughts):

- validation - very important for users and developers!
- modification of certain fields, e.g. translation between units
- analysis, e.g. extracting the building materials and generating a html
table becomes a trivial task with some xslt
- sometimes people may even want to open it in an editor (to fix a
broken file?)
- having it as a standard encoding to transfer between hosts (one host
may still have the data in a database, the other in one file, the next
in a file hierarchie, the next one generating it on the fly
(...)

Note that we should be more careful talking about the structure of data
or about how it is written somewhere. I am by far less interested in the
second. If you want, piping xml through bzip makes it a binary format...

> Like me. All people inside this list have the same problem. We need find
> solution to that...

This may become a biiig GNU project.

CU Lars.

[Lars O. Grobe]

A study on compression of xml formats and the discussion binary/text in
xml-land:

http://www.chris-augeri.com/docs/xml_compress.htm
http://www.chris-augeri.com/docs/pubs_files/ExpCS_Augeri_XML_corrected.pdf

(Links taken from today's oasis xml newsletter)

[martin_gnu]

On 28 sep, 08:40, "Lars O. Grobe" <gr...@gmx.net> wrote:
> A study on compression of xml formats and the discussion binary/text in
> xml-land:
>
> http://www.chris-augeri.com/docs/xml_compress.htm

I put this link here:
http://groups.google.com/group/open_cad_format/web/open-cad-format

All people can edit google pages also.

Regards.