primax: Thread: HELP WANTED


[<<] [<] Page 1 of 1 [>] [>>]
Subject: HELP WANTED
From: Andre Herms ####@####.####
Date: 31 Aug 2000 10:19:44 -0000
Message-Id: <00083112213501.01206@wh2-229>

Hi

Im writing the sane backend for Primax scanners.
And I need some help.

A short desciption waht is planed:
The sane backend is not the scanner driver, its just a client.
The server talks to the scanner.

The server:
It reads the data from the scanner and writes ist in raw format to a file.
the client can read it from there.

For making it work, there are some thing to do:

1. clean up the primaxscan code. the scanner control code must be seperated
from the output code (tiff/raw).

2. a clean makefile is needed for compiling the sane backend. I'm not good in
writing makefiles. Who can help?

3. The server shall read the settings from a config file and save them there. 
I think this is quiet easy to do with fprintf() and fscanf(). Could anyone
write some functions for this.
settings to save and read are:
int timeout
int warmup
int speed
int nice_level
char[100] password    // for accessing the advanced options of the server
calibration data      // this maybe better n another file.

If anyone reading this has enough time to do it....


PS: I hope the source code will reach a ftp or cvs server.

Bye 

  Andre

-- 
Andre Herms                   ####@####.####
                              http://www.cs.uni-magdeburg.de/~aherms
Subject: Re: HELP WANTED
From: Christian Biesinger ####@####.####
Date: 31 Aug 2000 11:31:21 -0000
Message-Id: <39AE4284.19497D6E@biesinger.net>

Andre Herms schrieb:
> 
> Hi
> 
> Im writing the sane backend for Primax scanners.
> And I need some help.

I'd like to help you.

> A short desciption waht is planed:
> The sane backend is not the scanner driver, its just a client.
> The server talks to the scanner.
> 
> The server:
> It reads the data from the scanner and writes ist in raw format to a file.
> the client can read it from there.

How does the client communicate with the server?
TCP/IP?
If so, why is the raw data not sent back using TCP/IP? Don't you think
that would be better?

> 
> For making it work, there are some thing to do:
> 
> 1. clean up the primaxscan code. the scanner control code must be seperated
> >from the output code (tiff/raw).

Probably won't be too difficult.
 
> 2. a clean makefile is needed for compiling the sane backend. I'm not good in
> writing makefiles. Who can help?

I think I could do this.

> 3. The server shall read the settings from a config file and save them there.
> I think this is quiet easy to do with fprintf() and fscanf(). Could anyone
> write some functions for this.
> settings to save and read are:
> int timeout
> int warmup
> int speed
> int nice_level
> char[100] password    // for accessing the advanced options of the server
> calibration data      // this maybe better n another file.

I'd like to do this as well, though I've got no idea what you mean with
calibration data and nice_level.


-- 
WAR IS PEACE - FREEDOM IS SLAVERY - IGNORANCE IS STRENGTH
2 x 2 = 5
Who controls the past controls the future:
Who controls the present controls the past
                              -- 1984, George Orwell
Subject: Re: HELP WANTED
From: Andre Herms ####@####.####
Date: 31 Aug 2000 12:52:42 -0000
Message-Id: <00083114542900.01681@wh2-229>

Hi
 
> How does the client communicate with the server?
> TCP/IP?
> If so, why is the raw data not sent back using TCP/IP? Don't you think
> that would be better?

I use pipes. Why should I use TCP/IP. I don't want to do network stuff.
Pipes has the advantage, that I can recognize immediately if the client crashes
or exits in an unusual way. Then the server can reset the connection and is
ready.

> > calibration data      // this maybe better n another file.
> 
> I'd like to do this as well, though I've got no idea what you mean with
> calibration data and nice_level.

calibration data: when the scan starts, it calibrates the color by getting the
white on the plastic. That's what always last so long befor the scanner starts
to scan. As it should normally not change, it should be enough to to it once
and save the result to a file.

nice level: unix is a timesharing system, the process with a higher nice level
gets mor cpu power. Since the scan can be time critical, it should be wise to
give the server on slower machines a higher level.
see also : "man nice"

-- 
Andre Herms                   ####@####.####
                              http://www.cs.uni-magdeburg.de/~aherms
Subject: Re: HELP WANTED
From: C Hennessy ####@####.####
Date: 31 Aug 2000 18:40:16 -0000
Message-Id: <39AEA6B3.D1E9E03@iname.com>

Christian how about creating a branch in the current cvs repository for this ?

Andre Herms wrote:
> 
> Hi
> 
> Im writing the sane backend for Primax scanners.
> And I need some help.
> 
> A short desciption waht is planed:
> The sane backend is not the scanner driver, its just a client.
> The server talks to the scanner.
> 
> The server:
> It reads the data from the scanner and writes ist in raw format to a file.
> the client can read it from there.
If I understand sane, this is one of the ways of communicating with it ?

I think to your list you could also add filling in the necessary data structures
in sane so that the sane frontend can access the necessary controls.

It would also be good to get a good idea from other people out there who
have taken a copy of the primaxsane code so that the changes made are compatabile
with what they have done.

CP
Subject: Re: HELP WANTED
From: ####@####.#### (Thomas Schano)
Date: 1 Sep 2000 23:29:51 -0000
Message-Id: <39B037DB.9FD9B539@t-online.de>

Andre Herms wrote:

> Hi
>
> Im writing the sane backend for Primax scanners.
> And I need some help.
>
> A short desciption waht is planed:
> The sane backend is not the scanner driver, its just a client.
> The server talks to the scanner.
>
> The server:
> It reads the data from the scanner and writes ist in raw format to a file.
> the client can read it from there.

That's described in the sane-spec

>
>
> For making it work, there are some thing to do:
>
> 1. clean up the primaxscan code. the scanner control code must be seperated
> from the output code (tiff/raw).

That's a very little step. It's no problem. The code should be proper almost. A
raw-data-output is only placing a case statement at two point in the code.

>
>
> 2. a clean makefile is needed for compiling the sane backend. I'm not good in
> writing makefiles. Who can help?

For sane a sane-makefile should be used. A copy of other makefiles is in the
sane-package. Such one could be modified.

>
> 3. The server shall read the settings from a config file and save them there.
> I think this is quiet easy to do with fprintf() and fscanf(). Could anyone
> write some functions for this.

Here is the main problem: There are 16 speeds the carriage can run. This would
mean to decide what settings are needed. Which format for the data sets should
be used an how could the records be separated? (I'm not a friend of binary files
for configuration.)

>
> settings to save and read are:
> int timeout

why this parameter?

>
> int warmup

why this one too?

>
> int speed

The speed should be choosen in dependency of the pixels per line.

>
> int nice_level

What's this good for? (I know the function of nice. But in my opinion it makes
no sense to change the priority of a driver.)

>
> char[100] password    // for accessing the advanced options of the server

I think a password should not be necessary. 100 Characters is a paragraph not a
word.

>
> calibration data      // this maybe better n another file.

Or 16 ones. Here is the main information inside.

>
>
> If anyone reading this has enough time to do it....

That's the main problem.

>
>
> PS: I hope the source code will reach a ftp or cvs server.

If you are writing it I'm sure it will.
....

Thomas Schano

PS. In my opinion porting the code to the kernel first would make more sense.
The 2nd step is the sane-port.



Subject: Re: HELP WANTED
From: Christian Biesinger ####@####.####
Date: 2 Sep 2000 08:58:19 -0000
Message-Id: <39B0C1AA.F3AD0DB2@biesinger.net>

Hello!

Thomas Schano schrieb:
> Andre Herms wrote:
> > A short desciption waht is planed:
> > The sane backend is not the scanner driver, its just a client.
> > The server talks to the scanner.
> >
> > The server:
> > It reads the data from the scanner and writes ist in raw format to a file.
> > the client can read it from there.
> 
> That's described in the sane-spec

?
It certainly isn't, because it has nothing to do with sane.

> > 3. The server shall read the settings from a config file and save them there.
> > I think this is quiet easy to do with fprintf() and fscanf(). Could anyone
> > write some functions for this.
> 
> Here is the main problem: There are 16 speeds the carriage can run. This would
> mean to decide what settings are needed. Which format for the data sets should
> be used an how could the records be separated? (I'm not a friend of binary files
> for configuration.)

I would use a textfile for it, and actually I'm writing some functions
to read and write the configuration.

> > int warmup
> 
> why this one too?

I can't exactly tell you why, but the scanner needs some time after the
light is switched on. The Windows driver needs this as well.

> > int nice_level
> 
> What's this good for? (I know the function of nice. But in my opinion it makes
> no sense to change the priority of a driver.)

A quote from another mail from Andre:
<begin quote>
nice level: unix is a timesharing system, the process with a higher nice
level
gets mor cpu power. Since the scan can be time critical, it should be
wise to
give the server on slower machines a higher level.
see also : "man nice"
<end quote>

> > PS: I hope the source code will reach a ftp or cvs server.
> 
> If you are writing it I'm sure it will.
> ....

Andre sent me the source code, should I upload it somewhere?

> Thomas Schano
> 
> PS. In my opinion porting the code to the kernel first would make more sense.
> The 2nd step is the sane-port.

Kernel?
The Linux Kernel?


-- 
WAR IS PEACE - FREEDOM IS SLAVERY - IGNORANCE IS STRENGTH
2 x 2 = 5
Who controls the past controls the future:
Who controls the present controls the past
                              -- 1984, George Orwell
Subject: Re: HELP WANTED
From: Simo Sorce ####@####.####
Date: 2 Sep 2000 09:00:51 -0000
Message-Id: <39B0C211.1FCCE308@tiscalinet.it>

Thomas Schano wrote:
> 
> Andre Herms wrote:
> 

...snip...

> >
> >
> > PS: I hope the source code will reach a ftp or cvs server.
> 
> If you are writing it I'm sure it will.
> ....
> 
> Thomas Schano
> 
> PS. In my opinion porting the code to the kernel first would make more sense.
> The 2nd step is the sane-port.


My kernel module is not finished nor has reached even the 5% of its
completenes.
But it already loads, detect the scanner and is able to move the
carriage.

It may however be a good framework to go on coding, if anyone is
interested in developing, I may make it available for download!

Simo.

-- 
####@####.####
http://www.geocities.com/SiliconValley/9757
Subject: Re: HELP WANTED
From: Puthukulangara Jaykrishnan ####@####.####
Date: 10 Jun 2001 07:24:56 -0000
Message-Id: <3B230BFC.814DCEC9@emirates.net.ae>

Dear Sir,
you are kindly requested to forward me the driver installation files
for  the scanner Primax Colorado direct 9600 to the below given URL : or
instruct us the procedure to download from the net.

####@####.####

your quick action will be highly appreciated.

best regards,
sajeev




Subject: Re: HELP WANTED
From: Christian Ordig ####@####.####
Date: 12 Jun 2001 13:35:12 -0000
Message-Id: <20010612153107.C8641@odin.chris.com>

On Sun, Jun 10, 2001 at 11:26:12AM +0530, Puthukulangara Jaykrishnan wrote:
> for  the scanner Primax Colorado direct 9600 to the below given URL : or
the latest version is downloadable via anonymous CVS... for details look
at our homepage.

-- 
Christian Ordig
Germany
[<<] [<] Page 1 of 1 [>] [>>]


Powered by ezmlm-browse 0.20.