Technical documentation scans to PDF

[dilinger]

Good evening,

 

Just a quick note about the scaned technical documentation, the ones found in the PDF,

I have done an OCR action on one of them (Technical Overview) to see the feasability to have a clean PDF file. I wanted to give you the result of this action.

What do you think about the results?

 

Thanks,

 

02-TechnicalOverview_10-18.April.1995.pdf

[JagChris]

Promising. This is from those scans?

[dilinger]

Yes.

I have used the PDF files which contains the scans provided by Lars Hannig.

I did some proof reading, but cannot guarantee the OCR is perfect.

[Clint Thompson]

I think it would be cool to have an online Jaguar resource center (website) for all documents and scans that would allow you to flip or page through it like a magazine, especially since we're on the whole touch-screen wave of things. Downloading is fine but with it being on-line all the time, it would or could always be there and easily accessible. Anyway, it looks good from what I can see... seems like a little updating and editiong could make it really shine.

[dilinger]

Thank you for the feedbacks,

I'm able to create a word file (by page), so I guess there is a way to improve the presentation, to have a unique good looking.

If you have to choose on future OCR actions, what chapter the community would like to see first? among the remaining ones:

Index.pdf

Getting Started.pdf

Software Reference.pdf

Technical Reference.pdf

Hardware Bugs & Warnings.pdf

Jaguar CD-ROM.pdf

The Jaguar Voice Modem.pdf

Jaguar Workshop Series.pdf

Sample Programs.pdf

Libraries.pdf

QSound for Jaguar.pdf

Cinepak for Jaguar.pdf

Tools.pdf

Appendices.pdf

Madmac Macro Assembler.pdf

ALN Linker.pdf

DB - The Atari Debugger.pdf

Thanks,

Edited December 6, 2014 by dilinger

[JagChris]

If you're going to do it then start from the beginning and let 'r rip.

[dilinger]

Ok, I do not know how far I will go but let's try it.

So, the next batch will be about the "Getting Started.pdf" file. I will keep you updated for this one.

[dilinger]

Performed OCR, and tried another presentation, on a smaller document "Hardware Bugs & Warnings"; it looks better now.

05-HardwareBugs&Warnings_26.April.1995.pdf

[+CyranoJ]

The Technical Reference manual has already been fully converted to OCR.

[+Stephen Moss]

I have been transcribing and correcting/updating the documents on and off for a couple of years, finished documents (off the top of my head) are...

Getting Started

Technical Reference

QSound

 

Nearly finished are...

Software Reference

Hardware Bugs & Warning (being incorported into the relevant parts of the other documents so you can say read about problems with the Blitter in the Blitter section instead of having to dive off into a different document).

 

Still have about 30 pages to go on the Software Reference and will need a few question answering to help clarify a couple of areas and then I will release them as a bundle once the software reference is finished.

 

Index is on going as I get the documents finished

 

Anyone with practical experience of working with the Jaguar who knows of any undocumented errors in the documents that need to be correct should PM me and I will take a look at changing the document/adding a note to it as appropriate.

[Zerosquare]

The Technical Reference manual has already been fully converted to OCR.

Yup, but it's actually an earlier revision. The scanned version is more recent and includes some corrections and additional info.

 

OCRing the scans is a really good idea. Even if the OCR cannot be trusted 100% (I'd recommend keeping the original pictures, and adding a "hidden" text layer, because human character recognition is better than OCR ), it makes the files searchable, which is great.

Edited December 9, 2014 by Zerosquare

[dilinger]

Thank you for the feedbacks,

 

I will try to finish the OCRs before Christmas; Since my Jaguar stuff is not arrived yet, I can spend some time on this,

 

This evening, I give you a new chapter "Getting Started"; as usual, feel free to let me know if you spot errors. It can be easily fixed. Since I'm new in the Jaguar scene, I use this OCR actions also to know more about the machine & the stuff around. So I apply some kind of proof reading.

 

The next chapter in progress is the "Technical Reference".

 

 

01-GettingStarted_10.April-5.June.1995.pdf

[ST-Oldie]

Hi,

 

I have done an OCR action on one of them (Technical Overview) ...

 

Nice to have the documents available as a text instead of pictures. This is much better. But did you know my package at http://www.mbernstein.de/download/jaguar/jdm.lzh? I have done a OCR over this documents in the past. I had send the package also to the guy who published the scaned Manuals. It seems he did not make much use of my work. The difference to your work is the file format. I had convert the files to the UDO format which is a meta format and can be converted to many other formats. Maybe this will help you to avoid some work.

 

Best regards

Michael

[dilinger]

Hello Michael,

 

Thank you very much for this information. Your work was unknown to me and I'm glad to know someone did OCR actions before.

 

I've tried UDO with your files but was not able to generate PDF or Rtf files. I'm using UDO 7.03 on Windows (command line) and I just got errors whatever the formats selected.

Most of the errors are related to "too many words..." or "items ignored...", but please could you generate one or two PDF and/or Rtf files at your side? To see how it looks like.

If you can also send me your options used in the command line, it will be useful, just in case of.

 

Anyway, your work offer a very interesting option / alternative for the community. And, I'm probably too old fashioned when it is about documentation - PDF or Word files have my preferences.

 

With kind regards,

[+remowilliams]

OCRing the scans is a really good idea. Even if the OCR cannot be trusted 100% (I'd recommend keeping the original pictures, and adding a "hidden" text layer, because human character recognition is better than OCR ), it makes the files searchable, which is great.

 

As someone who has done tons of documentation scans, this is why I very rarely use technologies like Clearscan when doing OCR...

[ST-Oldie]

Hi,

Thank you very much for this information. Your work was unknown to me and I'm glad to know someone did OCR actions before.

My work was not announced in any forum because at the time i download the scan and made ma OCR someone claims the right for publish the scans. And over the time i forgot to make this more public.

I've tried UDO with your files but was not able to generate PDF or Rtf files. I'm using UDO 7.03 on Windows (command line) and I just got errors whatever the formats selected.
Most of the errors are related to "too many words..." or "items ignored...", but please could you generate one or two PDF and/or Rtf files at your side? To see how it looks like.
If you can also send me your options used in the command line, it will be useful, just in case of.

 

I use udo 6.4 on Atari, but with UDO-Shell. And i use the defaults from UDO-Shell. I dont know which parameters tis gui use. I did not get this problems. I can check with a newer version of UDO, maybe on a different system. I made a rtf with my UDO system, but without images, you can get it at http://www.mbernstein.de/download/jaguar/jdm.zip The output is different from your work because you try to rebuild the original format, but as text instead of a image. UDO did not take care of this, so it looks different.

 

And, I'm probably too old fashioned when it is about documentation - PDF or Word files have my preferences.

The same for me. But because i started with Atari Computers, some old Atari document formats are more common to me (at this time). PDF and RTF are to modern and not so common. PDF always did not work as well than on PC. But ST-Guide help system and html was much more interesteing. HTML also for online.

Best regards
Michael

[dilinger]

Hi Michael,

 

Thank you for the file, I have look at it and it looks fine.

Look / presentation is different but this is a matter of taste, the information are here and this is the most important.

 

Anyway, your version is already available so the community has a nice alternative; I wll continue to do the PDF and will post them here time to time.

 

Regards,

04-TechnicalReference_26.April-21.June.1995.pdf

[ST-Oldie]

Hello,

 

I've tried UDO with your files but was not able to generate PDF or Rtf files. I'm using UDO 7.03 on Windows (command line) and I just got errors whatever the formats selected.

Most of the errors are related to "too many words..." or "items ignored...", but please could you generate one or two PDF and/or Rtf files at your side? To see how it looks like.

If you can also send me your options used in the command line, it will be useful, just in case of.

I made a test with UDO 7.03 on Windows and got one error and UDO stops to convert the text. After some digging i found 2 bugs with begin_node/end_node (chapter depth). It seems, the handling of errors is different in UDO 6 which i have used in the past and the actual UDO 7. I made a bugfix in the UDO source files. Now it converts fine. I got many warnings withj UDO 7 about "short is deprecated" but this can be ignored. I made my conversion to a test rtf without bitmaps. So i got also error about missing bmp files.

 

To convert the udo source, first the right pictures should be copied into the folder with the udo source. For rtf the pictures must also converted to windows bmp format.

 

I created the udo source on Atari ST. If the udo source is converted on an other platform than TOS, a "!code_source TOS" may be a good idea. This leads UDO to expect Atari ST Codepage for the input source.

 

I did not see any other errors or warning with my test. Maybe there is a problem with the case of the filenames?

 

I made an update of the source and also of the test rtf on my homepage.

 

Best regards

Michael

[dilinger]

Hi Michael,

 

Thank you for the updates, and sorry for this late reply.

 

I have tried with your new package and just got 2 errors (on the ALN.UI file by example), errors different from the previous ones.

"maximum token number exceeded without further..."

 

I will look at the issue again, but your package looks very usable anyway.

 

Regards

[ST-Oldie]

Hi

Thank you for the updates, and sorry for this late reply.

No problem, we have no time pressure.

I have tried with your new package and just got 2 errors (on the ALN.UI file by example), errors different from the previous ones.

"maximum token number exceeded without further..."

Maybe i did not see this error because of the warnings about "!short". I dont know if i have time left before christmas to take a look at this. If you like to look for it, feel free to do this.

 

Best regards

Michael