Keeping Rawnalyze alive, a tribute to GaborSch

Started Aug 22, 2010 | Discussions
Joseph S Wisniewski Forum Pro • Posts: 34,130
Keeping Rawnalyze alive, a tribute to GaborSch

For those of you who don't know already, GaborSch, one of the most knowledgeable, courteous, and sharing people on dpReview passed away recently. Here's a little bit about what happened, and some nice tributes...

http://forums.dpreview.com/forums/read.asp?forum=1032&message=36006562

Aside from copious amounts of excellent advice, freely given, Gabor also gave us a freeware raw file analysis program called Rawnalyze.

There has been discussion recently about trying to continue developing Rawnalyze. I'm starting a new thread about that, here, because the original thread is in the Canon forum, and I think it's more an "Open Talk" thing, and because I'd like to not have the "say nice things about Gabor" thread turn into a thread about software development.

Although Gabor gave Rawnalyze out as freeware, he did not open the source. That's quite common: for many of us, programming is highly personal: it gives insights into our thought processes that we might not want to share. I know I don't like people looking at the code for my large, personal projects like mImage or WordWhacker.

Right now, there's a discussion going on about how to keep Rawnalyze alive. To continue developing Gabor's program would involve three things.

  • Getting access to his source.

  • Assembling his development environment (I have no idea what compiler or libraries he used, or how portable it is).

  • Figuring out enough about how it works to be able to expand on it.

To develop a "Rawnalyze inspired" program would be a little more difficult.

  • Come up with some sort of guiding document.

  • Choose a development environment and libraries that support a "group think" style of work.

I'm thinking of trying the second approach. Gabor once mentioned, in a conversation involving a Mac version of Rawnalyze, that it wasn't sufficiently portable. I've recently begun playing with Iliah Borg's "libraw", which is a portable, thread-safe, callable library "reimaging" of Dave Coffin's dcraw. Additionally, it does some things with black levels and dark frames that are very, very useful for a project of this type. And I always build these sort of programs with wxWidgets (which I've been using for about 10 years now), and I have need of a wx wrapper around LibRaw for entirely different purposes, but would make that wrapper open source as it grows. I tinkered last night, and it took me about an hour to get something that called LibRaw from wx, and could at least get the dimensions of raw files.

So, here's my thoughts, so far.

  • Project hosting: Sourceforge.

  • Project name: Rawnalyze, RawGabor, Sch, SchRaw. Open to suggestions.

  • Application framework: wxWidgets 2.9.1. A 2.9 fork leads directly to wx 3, which is supposed to be out this year.

  • Raw access: LibRaw. It doesn't seem to be that hard, and for this project, it's better than the way mImage uses worker threads to exec dcraw.

  • Gamma curves and viewing gamuts: Little CMS 2. I haven't used LCMS 2 yet, but I've been looking forward to it, and it's portable.

  • Version control: open to suggestions.

And some future directions

  • Database: not sure if it needs one, but if it does, same as mImage: SQLite with the wxSQLite3 wrapper. Schemas overlapping Lightroom 3.

  • i18n: not my strong point, open to suggestions. Poedit?

  • Mac port: wx makes this pretty easy. Anyone have a Mac and a compiler?

Suggestions, catcalls, etc?

Oh, and if there's anyone who has Rawnalyze 2.10.0.4 and would like to either send me a copy, or summarize what's changed from 2.9.7.0, it would be much appreciated.

-- hide signature --

Rahon Klavanian 1912-2008.

Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed.

Ciao! Joseph

http://www.swissarmyfork.com

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Nikon D3 Nikon D2X Nikon D90 Nikon D100 Canon EOS 5D Mark II +43 more
Cedarhill Veteran Member • Posts: 4,638
Re: Keeping Rawnalyze alive, a tribute to GaborSch

Bravo! I am not familiar with his comments and I haven't tried his program but I think it is great that you are hoping to carry it forward. I would suggest using his name in the new project as a kind of tribute.

bronxbombers Forum Pro • Posts: 18,226
Re: Keeping Rawnalyze alive, a tribute to GaborSch

Magne Nilsen wrote:
bronxbombers wrote:
Per Inge Oestmoen wrote:
Anonimo wrote:
Hi Joseph!
Thanks for that.

I understood that somebody in the forum wanted to contact Gabor's family to send our condolences. Maybe the person can tactfully ask if they have any plans for Rawanalyze (it is Gabor's intelectual property) and if they would agree that someone keep it alive. We could use the community to search for a suitable candidate.

One possible way would be to offer the code to some open-source projects. Maybe Dave Coffin (dcraw's author) or the LibRAW crew gets interested.

Hello!

Judging from Gabor's noble personality, high intelligence and willingness to help, I would not be surprised if it was his wish to let Rawnalyze (which is the correct name) go Open Source.

Per Inge Oestmoen, Norway

perhaps so (although he had been extremely secretive about the code when he was working on it, making special private internal versions, etc. and treating it as 100% opposite of open source; all the same would he actually want it to just disappear??)

hmmm...

from my correspondence with Gabor regarding coding in general, but also related to specific programming projects, I think that he looked upon his programming as a rather private thing.

My personal feeling would be that he would not have preferred his source code to be perused by others. I could be wrong under the circumstances, but I saw no reasons to ask twice back then. As usual, he was rather clear.
-- hide signature --

Magne

Perhaps I was trying to be a bit too optimistic there and trying to agree to what extent I could with Per Inge Oestmoen. He did seem to be extremely secretive about his coding, for whatever reason, and even kept secret versions of even the compiled version of Raw only for himself. He didn't sound too open at all when I asked about being to code anything for it and I sense that I should quickly drop the subject. So I'm not sure what to say.

I should add to my response above though that it is very common to not want to have someone messing with code you are still working on, many times once you no longer work or bother with some old code you have totally abandoned you might likely suddenly not care at all who uses it.

Great Bustard Forum Pro • Posts: 40,694
Suggestions, catcalls, etc.

Joseph S Wisniewski wrote:

Suggestions, catcalls, etc?

I think it's important to contact the family, express the intent, and abide by their wishes. That said, it seems to me that to keep his work alive is a wonderful tribute to his contributions, and I commend those that would seek to keep his legacy alive.

OP Joseph S Wisniewski Forum Pro • Posts: 34,130
Contact the family...

Great Bustard wrote:

Joseph S Wisniewski wrote:

Suggestions, catcalls, etc?

I think it's important to contact the family, express the intent, and abide by their wishes. That said, it seems to me that to keep his work alive is a wonderful tribute to his contributions, and I commend those that would seek to keep his legacy alive.

Excellent suggestion.

Now, if someone who has the contact info will forward it to me...

-- hide signature --

Rahon Klavanian 1912-2008.

Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed.

Ciao! Joseph

http://www.swissarmyfork.com

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Nikon D3 Nikon D2X Nikon D90 Nikon D100 Canon EOS 5D Mark II +43 more
J A K Forum Pro • Posts: 15,833
Joe

Both Andre and Victor (profiles below) have contacted the family.

http://www.dpreview.com/members/8769164387

http://www.dpreview.com/members/9533184395

FWIW I have version 2.9.9.5; please let me know if that version will be of any help to you.

BTW, very smart move on your part to move the RawAnalyze discussion to the Open Forum.

Regards,

Joe Kurkjian

Galleries: http://www.pbase.com/jkurkjia

-- hide signature --

SEARCHING FOR A BETTER SELF PORTRAIT

Daniel Browning Senior Member • Posts: 1,058
How about as a modification to RawTherapee?

Joseph S Wisniewski wrote:

I'm thinking of trying the second approach.

That would be great. My C is a bit rusty, but I would like to be able to contribute too.

  • Project hosting: Sourceforge.

  • Project name: Rawnalyze, RawGabor, Sch, SchRaw. Open to suggestions.

I like SchRaw. Other suggestions: RawnalyzeB, Brawnalyze, Rawspector, Rawspect, Rawloupe.

  • Application framework: wxWidgets 2.9.1. A 2.9 fork leads directly to wx 3, which is supposed to be out this year.

Great idea for cross-platform compatibility.

  • Raw access: LibRaw. It doesn't seem to be that hard, and for this project, it's better than the way mImage uses worker threads to exec dcraw.

That way would be smart, too.

  • Version control: open to suggestions.

Personally, I would prefer Git or Mercurial.

And some future directions

  • Mac port: wx makes this pretty easy. Anyone have a Mac and a compiler?

"Not it!"

You know, another possibility would be to build this into RawTherapee directly. One method might be as an alternative demosaic (with rawnalyze-style options).
--
Daniel

Per Inge Oestmoen Senior Member • Posts: 1,759
Re: Keeping Rawnalyze alive, a tribute to GaborSch

bronxbombers wrote:

Perhaps I was trying to be a bit too optimistic there and trying to agree to what extent I could with Per Inge Oestmoen. He did seem to be extremely secretive about his coding, for whatever reason, and even kept secret versions of even the compiled version of Raw only for himself. He didn't sound too open at all when I asked about being to code anything for it and I sense that I should quickly drop the subject. So I'm not sure what to say.

I should add to my response above though that it is very common to not want to have someone messing with code you are still working on, many times once you no longer work or bother with some old code you have totally abandoned you might likely suddenly not care at all who uses it.

My latest comment in that thread, for the full context:

My intuitive feeling from my correspondence with Gabor was that he viewed Rawnalyze as very private, but not in the sense that he showed the sometimes neurotic approach to "intellectual property" which expresses itself in such undesirable phenomena as Product Activation with ties to a manufacturer after the licensee has paid the fee. Gabor obviously had very high standards for himself and, in my understanding, simply wanted to ensure that the program maintained the quality he demanded by doing it all by himself. I feel that were this not so, he would not have given his program away from free.

My take on this question is that it is as likely as not that he would have wanted his program to become Open Source instead of dying with him - but he would also be VERY anxious to know that his standards were adhered to.

Per Inge Oestmoen, Norway

  • I am not surprised at all that he wanted to keep his code secret. It is a normal thing to be very protective of your creation; be it a child or the fruit of some other kind of artistic or creative activity. But with his passing away, Gabor's situation has changed and the question is what would have been his wish now. He cannot participate in this discussion or communicate with us now, but there was a reason why he decided to let Rawnalyze be a freeware program and let it be available to all who could benefit from it.

Has anyone received input from his family?

Per Inge Oestmoen, Norwy

Dave_Anderson Senior Member • Posts: 1,471
Very sorry to hear of this.

I did not know Gabor but found Rawnalyze useful. Clearly there is a need that has to date only been filled by Rawnalyze AFAIK.

Much of the site information is still available here: http://web.archive.org/web/*/http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

The download links do not work, unfortunately.

I did not get v2.10.0.4 but I do have v2.10.1.0 for PC. Not sure why you would want an earlier version. Before I let it go I will post an MD5 here so that others may cross-check and ensure that the other copies that are sure to crop up are unaltered.

Dave_Anderson Senior Member • Posts: 1,471
Re: Very sorry to hear of this.

Dave_Anderson wrote:

I did not know Gabor but found Rawnalyze useful. Clearly there is a need that has to date only been filled by Rawnalyze AFAIK.

Much of the site information is still available here: http://web.archive.org/web/*/http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

The download links do not work, unfortunately.

I did not get v2.10.0.4 but I do have v2.10.1.0 for PC. Not sure why you would want an earlier version. Before I let it go I will post an MD5 here so that others may cross-check and ensure that the other copies that are sure to crop up are unaltered.

Here it is:

The .exe is at the same location as the .md5 file, just change the extension to .exe
http://dave-anderson-photo.com/files/software/Rawnalyze/Rawnalyze_2.10.1.0.md5

$ md5sum -b Rawnalyze_2.10.1.0.exe
b94df7febb61857ed2fc85260f99920e Rawnalyze_2.10.1.0.exe
$

Joofa Senior Member • Posts: 2,564
Re: Keeping Rawnalyze alive, a tribute to GaborSch

Joseph S Wisniewski wrote:

  • Mac port: wx makes this pretty easy. Anyone have a Mac and a compiler?

If the project is setup right with makefiles, as on a linux environment, then it can be a good starting point for the Mac environment.

Joofa

-- hide signature --

Dj Joofa

Andre Affleck Senior Member • Posts: 2,362
Re: Joe

Victor has indicated that he will take the initiative to contact the family regarding the fate of RawAnalyze and will post back on the original thread what the response is. Let's wait and see what Victor finds out. I think it would be best to have one point of contact at this time instead of having everyone inundate the family with requests.

J A K wrote:

Both Andre and Victor (profiles below) have contacted the family.

http://www.dpreview.com/members/8769164387

http://www.dpreview.com/members/9533184395

FWIW I have version 2.9.9.5; please let me know if that version will be of any help to you.

BTW, very smart move on your part to move the RawAnalyze discussion to the Open Forum.

Regards,

Joe Kurkjian

Galleries: http://www.pbase.com/jkurkjia

-- hide signature --

SEARCHING FOR A BETTER SELF PORTRAIT

OP Joseph S Wisniewski Forum Pro • Posts: 34,130
Me to. And thatnk you...

Dave_Anderson wrote:

I did not know Gabor but found Rawnalyze useful. Clearly there is a need that has to date only been filled by Rawnalyze AFAIK.

Much of the site information is still available here: http://web.archive.org/web/*/http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

The download links do not work, unfortunately.

I did not get v2.10.0.4 but I do have v2.10.1.0 for PC. Not sure why you would want an earlier version. Before I let it go I will post an MD5 here so that others may cross-check and ensure that the other copies that are sure to crop up are unaltered.

Thank you greatly.

I only mentioned 2.10.0.4. because I thought it was the latest version. Your 2.10.1.0 is ever better. Thank you again.

-- hide signature --

Rahon Klavanian 1912-2008.

Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed.

Ciao! Joseph

http://www.swissarmyfork.com

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Nikon D3 Nikon D2X Nikon D90 Nikon D100 Canon EOS 5D Mark II +43 more
OP Joseph S Wisniewski Forum Pro • Posts: 34,130
Works for me...

Andre Affleck wrote:

Victor has indicated that he will take the initiative to contact the family regarding the fate of RawAnalyze and will post back on the original thread what the response is.

OK. That works for me.

Let's wait and see what Victor finds out. I think it would be best to have one point of contact at this time instead of having everyone inundate the family with requests.

Definitely. Victor is welcomed to be Spock on this one.

J A K wrote:

Both Andre and Victor (profiles below) have contacted the family.

http://www.dpreview.com/members/8769164387

http://www.dpreview.com/members/9533184395

FWIW I have version 2.9.9.5; please let me know if that version will be of any help to you.

BTW, very smart move on your part to move the RawAnalyze discussion to the Open Forum.

-- hide signature --

Rahon Klavanian 1912-2008.

Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed.

Ciao! Joseph

http://www.swissarmyfork.com

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Nikon D3 Nikon D2X Nikon D90 Nikon D100 Canon EOS 5D Mark II +43 more
Dave_Anderson Senior Member • Posts: 1,471
Re: Me to. And thatnk you...

Joseph S Wisniewski wrote:

Dave_Anderson wrote:

I did not know Gabor but found Rawnalyze useful. Clearly there is a need that has to date only been filled by Rawnalyze AFAIK.

Much of the site information is still available here: http://web.archive.org/web/*/http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

The download links do not work, unfortunately.

I did not get v2.10.0.4 but I do have v2.10.1.0 for PC. Not sure why you would want an earlier version. Before I let it go I will post an MD5 here so that others may cross-check and ensure that the other copies that are sure to crop up are unaltered.

Thank you greatly.

I only mentioned 2.10.0.4. because I thought it was the latest version. Your 2.10.1.0 is ever better. Thank you again.

You're welcome. I wish you and all who are trying to move development forward the best of luck. Due to the nature of RAW this sort of software needs constant updates, and I hope that whoever takes on the role of Steward(assuming this happens) puts some contingency plans in place.

exdeejjjaaaa
exdeejjjaaaa Veteran Member • Posts: 8,263
Re: Me to. And thatnk you...

it might be better to join some existing project like rawtherapee ( http://www.rawtherapee.com/forum/index.php ) and develop some kind of raw analyzer module to be used there as an addon to the raw converter which rawtherapee is... advantage - you have already GUI working there.

-- hide signature --

 exdeejjjaaaa's gear list:exdeejjjaaaa's gear list
Sony Alpha a7R II Sony FE 55mm F1.8 Phase One Capture One Pro +25 more
TrapperJohn Forum Pro • Posts: 16,488
Well, it's a start

We have one project (a fairly large one) at my company written with wxWidgets. It sort of ports to other systems, but when one gets down and dirty, there are still some OS specific #ifdefs in your own code.

That was true of Win to Linux cross compiling, I didn't work with wxWidgets under Mac. Don't recall what they did with little/big endian, that wasn't an issue with Linux. As Mac moves away from PPC, it's becoming less of an issue there.

Having that, I'd love to see the source go to sourceforge, just to have a peek. My background is in color separation and calibration for printer firmware (4 color for laser, nothing fancy), but I've been looking for an excuse to push into more detailed areas.

Tyrone Wellhung Contributing Member • Posts: 980
Re: Keeping Rawnalyze alive, a tribute to GaborSch

Hi Joseph,

late into this thread, but starting it is an excellent idea. I'm doing some stuff at the moment which would be much easier with Rawnalyze, so I believe there is a clear need, and you've my full support on this. A few comments as appropriate below:

Joseph S Wisniewski wrote:

Right now, there's a discussion going on about how to keep Rawnalyze alive. To continue developing Gabor's program would involve three things.

  • Getting access to his source.

  • Assembling his development environment (I have no idea what compiler or libraries he used, or how portable it is).

  • Figuring out enough about how it works to be able to expand on it.

To develop a "Rawnalyze inspired" program would be a little more difficult.

I think the reverse. I wouldn't mind betting that Gabor's code is highly individual, just understanding it is likely to be a big job in itself. In the end, a new implementation would probably be easier and more reliable, and also easier to do as a collaborative project.

  • Come up with some sort of guiding document.

Maybe we need an open discussion on what people want to do with it, what's good and bad with Rawnalyze, and what other software and systems people want to use it with. Develop to a spec for what we want to do.

  • Choose a development environment and libraries that support a "group think" style of work.

There's a lot to be said for emulating existing such projects. Plenty of them on Sourceforge.

I'm thinking of trying the second approach.

I agree with that

I've recently begun playing with Iliah Borg's "libraw", which is a portable, thread-safe, callable library "reimaging" of Dave Coffin's dcraw. Additionally, it does some things with black levels and dark frames that are very, very useful for a project of this type.

I agree with that too.

And I always build these sort of programs with wxWidgets (which I've been using for about 10 years now), and I have need of a wx wrapper around LibRaw for entirely different purposes, but would make that wrapper open source as it grows. I tinkered last night, and it took me about an hour to get something that called LibRaw from wx, and could at least get the dimensions of raw files.

So, here's my thoughts, so far.

  • Project hosting: Sourceforge.

  • Project name: Rawnalyze, RawGabor, Sch, SchRaw. Open to suggestions.

I like SchRaw, but I also like messing with acronyms, so how about Generalized Analysis and Breakdown Of Raw?

  • Application framework: wxWidgets 2.9.1. A 2.9 fork leads directly to wx 3, which is supposed to be out this year.

  • Raw access: LibRaw. It doesn't seem to be that hard, and for this project, it's better than the way mImage uses worker threads to exec dcraw.

  • Gamma curves and viewing gamuts: Little CMS 2. I haven't used LCMS 2 yet, but I've been looking forward to it, and it's portable.

  • Version control: open to suggestions.

And some future directions

  • Database: not sure if it needs one, but if it does, same as mImage: SQLite with the wxSQLite3 wrapper. Schemas overlapping Lightroom 3.

I'm not sure it needs one at present, but I can see basing around a datbase will give it a lot more growth and power.

  • i18n: not my strong point, open to suggestions. Poedit?

  • Mac port: wx makes this pretty easy. Anyone have a Mac and a compiler?

We should aim for a neutral Unix version that would run OS-X or Linux (or any other Unix). Personally, I'd go for making this the master version but that's probably just me.

Suggestions, catcalls, etc?

At some time it ought to include a scripting language. I'd suggest Python, which seems to be the scripting language of choice.
Good call in any case, count me in.

Bob

Joofa Senior Member • Posts: 2,564
Re: Well, it's a start

TrapperJohn wrote:

That was true of Win to Linux cross compiling, I didn't work with wxWidgets under Mac. Don't recall what they did with little/big endian, that wasn't an issue with Linux. As Mac moves away from PPC, it's becoming less of an issue there.

Hi,

I seem to recall compling wxWidgets several years back on my PPC Mac laptop. It wasn't an easy effort. At that time several such programs had problems on Mac and I had to spend weeks figuring out the problems and changing code around to make them compile.

I have not tried it on the Intel based Mac though.

However, as I seem to recall some problems used to stem from issues such as difference between Linux and OS X version of libtool, etc.

Sincerely,

Joofa

-- hide signature --

Dj Joofa

Dave_Anderson Senior Member • Posts: 1,471
Re: Very sorry to hear of this.

Dave_Anderson wrote:

Dave_Anderson wrote:

I did not know Gabor but found Rawnalyze useful. Clearly there is a need that has to date only been filled by Rawnalyze AFAIK.

Much of the site information is still available here: http://web.archive.org/web/*/http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

The download links do not work, unfortunately.

I did not get v2.10.0.4 but I do have v2.10.1.0 for PC. Not sure why you would want an earlier version. Before I let it go I will post an MD5 here so that others may cross-check and ensure that the other copies that are sure to crop up are unaltered.

Here it is:

The .exe is at the same location as the .md5 file, just change the extension to .exe
http://dave-anderson-photo.com/files/software/Rawnalyze/Rawnalyze_2.10.1.0.md5

$ md5sum -b Rawnalyze_2.10.1.0.exe
b94df7febb61857ed2fc85260f99920e Rawnalyze_2.10.1.0.exe
$

I have pulled the aforementioned .exe file from the server. I have replaced it with a zipfile that also contains the MD5 data and I have done MD5s of the zips. Additionally, I have uploaded a later version to my site. I downloaded 2.10.1.0 myself from Gabors site and I will vouch for that. The later version(2.10.3.0) came to me with very little information but passed a "sniff test" of sorts.

In the event an altered copy of one of these files turns up or in the unlikely event that one of my copies turns out to be altered MD5 is a convenient way to detect whether two files are truly identical. Please, if you have one of these versions and obtain a different MD5 from it, get in touch with me so we can figure out what's going on.

I have made a blog post at the following link with the pertinent details. I will post updates here as appropriate, but the blog post will always have the most up-to-date info(and links that work for people )

http://dave-anderson-photo.com/blog/2010/08/23/gabor-rawnalyze-author-rip/

Keyboard shortcuts:
FForum MMy threads