Understanding LRS format - Page 2

RWood · 09-12-2007, 06:01 PM

These are the reasons that I don't code anymore.

silkcom · 09-12-2007, 10:15 PM

It just seems to me that If I can create an editor that goes directly to LRS then I would be sure that it shows up the way I want it too. That's my reasoning. Then I can use lrs2lrf to convert it, and no problems with converting stuff. I've had my share of problems with the different python converters.

I've been looking through this LRS document, and it doesn't seem any worse than HTML?. And if I create a Java library for writing LRS then I know that it's the quality I would like.

kovidgoyal · 09-12-2007, 10:42 PM

LRS contains no semantic information. Even vanilla html contains header tags and tags like <em> <strong> <li> etc. And with CSS you can make HTML fully semantic. LRS on the other hand is purely layout oriented. And not very capable at that.

There is no cross platform implementation of lrs2lrf though making one shouldn't be difficult.

If you have problems with existing converters why not help improve them?

silkcom · 09-12-2007, 11:19 PM

I'm more than willing, but I'm really a Java man. I don't know python, and yours seems to be the best, except that it's in python. But I would be willing to help, how does one help with the libprs stuff?

On the other side though I would prefer for the whole thing to also be done in Java. I've seen that you've said that having both isn't a good idea, but I think that it brings more programmers.

kovidgoyal · 09-12-2007, 11:58 PM

Well far be it from me to stop you. Personally, I found python a remarkably easy language to pick up, so I encourage people to dive in. But then I'm still young enough to find learning a new language exciting :-)

If you want to help with libprs500, there are instructions at https://libprs500.kovidgoyal.net/wiki/Development

You can always contact me if you have questions.

silkcom · 09-13-2007, 11:29 AM

What are the chances of our being able to make the PDF to HTML more powerful? Is that a limitation with some library? It completely removes all forms of formatting

(not acceptable with the documents I'm reading rasterfarian works better).

I have your code, and have started to look through it. What development environment options do I have with python and windows?

kovidgoyal · 09-13-2007, 12:04 PM

libprs500 relies on the pdftohtml library to first convert pdf to html and then process it. I'm afraid that pdftohtml already works as well as Adobe's own PDF to HTML conversion, so I doubt it can be improved more without considerable effort, but you're welcome to try.

I'd recommend eclipse + pydev since you're coming from the java world.

silkcom · 09-13-2007, 01:56 PM

Anyone else willing to play with PDF to HTML to make it better?? It really needs to keep as much formatting as possible (without being ridiculous). So that the books can be converted and still look basically the same.

Significant effort, but when we're done we'll have an actual solution. As of now, the PDF to HTML is worthless to me.

ashkulz · 09-17-2007, 06:44 AM

Quote:

Originally Posted by kovidgoyal

Incidentally most of the LRS Spec, as well as the conversion from LRS->LRF (which is not documented by SONY) is defined through an object library pylrs, which is part of libprs500.

PyLRS is actually a separate library mainly developed by Falstaff and later on incorporated into many projects (libprs500 and pdfread being some examples). All of them have made some (relatively minor) tweaks to the original. Credit where credit is due

kovidgoyal · 09-17-2007, 12:56 PM

Sure, I didn't mean to steal credit. There's also an implemetation of lrf2lrs which was first developed by roxfan and igorsk but that is also maintained as part of libprs500 by me.

Adam B. · 09-17-2007, 05:10 PM

It would be nice if there was an LRS viewer written in C so that LRS formatted files could be viewed in Linux and on other mobile platforms.

kovidgoyal · 09-17-2007, 05:12 PM

I'm actually writing a LRF viewer, but it's going to be in python.

Adam B. · 09-17-2007, 05:29 PM

Quote:

Originally Posted by kovidgoyal

I'm actually writing a LRF viewer, but it's going to be in python.

Maybe that will be good enough... Python is at least partially work on the iLiad.

I'll keep an eye out for it. Thanks.

aapezzuto · 10-12-2007, 07:02 PM

is there a way to compile python into a native executable on platforms other than windows? I have only used python because the native scripting facilities in windows are to unintuitive, and pearl made me want to cry.
Long live the interpreted languages, strongly typed, and otherwise (I love java and python)

igorsk · 10-12-2007, 07:12 PM

You know a python-to-native compiler for windows? The closest I know is psyco and it's more like a JIT than a compiler.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Converting LRS to LRF	mcortez	Sony Reader	1	11-28-2007 12:26 PM
PRS-500 Create LRS and LRF files from Python with pylrs	Falstaff	Sony Reader Dev Corner	37	08-04-2007 08:49 PM
LRS to LRF on Mac OSX - any suggestions?	Lime2K	Sony Reader	3	07-29-2007 11:15 PM
PRS-500 Span tags in LRS and LRF files -- do I understand them?	Falstaff	Sony Reader Dev Corner	2	01-31-2007 11:34 AM
PRS-500 BBeB Xylog XML (LRS) final spec released	igorsk	Sony Reader Dev Corner	5	01-12-2007 03:04 AM

09-12-2007, 06:01 PM	#16
RWood Technogeezer Posts: 7,233 Karma: 1601464 Join Date: Nov 2006 Location: Virginia, USA Device: Sony PRS-500	These are the reasons that I don't code anymore.

09-12-2007, 10:15 PM	#17
silkcom Member Posts: 20 Karma: 10 Join Date: Sep 2007 Device: PRS-500	It just seems to me that If I can create an editor that goes directly to LRS then I would be sure that it shows up the way I want it too. That's my reasoning. Then I can use lrs2lrf to convert it, and no problems with converting stuff. I've had my share of problems with the different python converters. I've been looking through this LRS document, and it doesn't seem any worse than HTML?. And if I create a Java library for writing LRS then I know that it's the quality I would like.

09-12-2007, 10:42 PM	#18
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	LRS contains no semantic information. Even vanilla html contains header tags and tags like <em> <strong> <li> etc. And with CSS you can make HTML fully semantic. LRS on the other hand is purely layout oriented. And not very capable at that. There is no cross platform implementation of lrs2lrf though making one shouldn't be difficult. If you have problems with existing converters why not help improve them?

09-12-2007, 11:19 PM	#19
silkcom Member Posts: 20 Karma: 10 Join Date: Sep 2007 Device: PRS-500	I'm more than willing, but I'm really a Java man. I don't know python, and yours seems to be the best, except that it's in python. But I would be willing to help, how does one help with the libprs stuff? On the other side though I would prefer for the whole thing to also be done in Java. I've seen that you've said that having both isn't a good idea, but I think that it brings more programmers.

09-12-2007, 11:58 PM	#20
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Well far be it from me to stop you. Personally, I found python a remarkably easy language to pick up, so I encourage people to dive in. But then I'm still young enough to find learning a new language exciting :-) If you want to help with libprs500, there are instructions at https://libprs500.kovidgoyal.net/wiki/Development You can always contact me if you have questions.

09-13-2007, 11:29 AM	#21
silkcom Member Posts: 20 Karma: 10 Join Date: Sep 2007 Device: PRS-500	What are the chances of our being able to make the PDF to HTML more powerful? Is that a limitation with some library? It completely removes all forms of formatting (not acceptable with the documents I'm reading rasterfarian works better). I have your code, and have started to look through it. What development environment options do I have with python and windows?

09-13-2007, 12:04 PM	#22
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	libprs500 relies on the pdftohtml library to first convert pdf to html and then process it. I'm afraid that pdftohtml already works as well as Adobe's own PDF to HTML conversion, so I doubt it can be improved more without considerable effort, but you're welcome to try. I'd recommend eclipse + pydev since you're coming from the java world.

09-13-2007, 01:56 PM	#23
silkcom Member Posts: 20 Karma: 10 Join Date: Sep 2007 Device: PRS-500	Anyone else willing to play with PDF to HTML to make it better?? It really needs to keep as much formatting as possible (without being ridiculous). So that the books can be converted and still look basically the same. Significant effort, but when we're done we'll have an actual solution. As of now, the PDF to HTML is worthless to me.

09-17-2007, 12:56 PM	#25
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Sure, I didn't mean to steal credit. There's also an implemetation of lrf2lrs which was first developed by roxfan and igorsk but that is also maintained as part of libprs500 by me.

09-17-2007, 05:10 PM	#26
Adam B. Addicted to Porting Posts: 1,697 Karma: 7194 Join Date: Oct 2006 Location: Indianapolis, IN Device: iRex iLiad, Nokia 770, Samsung i760	It would be nice if there was an LRS viewer written in C so that LRS formatted files could be viewed in Linux and on other mobile platforms.

09-17-2007, 05:12 PM	#27
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	I'm actually writing a LRF viewer, but it's going to be in python.

10-12-2007, 07:02 PM	#29
aapezzuto Zealot Posts: 116 Karma: 693 Join Date: Sep 2007 Device: none	is there a way to compile python into a native executable on platforms other than windows? I have only used python because the native scripting facilities in windows are to unintuitive, and pearl made me want to cry. Long live the interpreted languages, strongly typed, and otherwise (I love java and python)

10-12-2007, 07:12 PM	#30
igorsk Wizard Posts: 3,442 Karma: 300001 Join Date: Sep 2006 Location: Belgium Device: PRS-500/505/700, Kindle, Cybook Gen3, Words Gear	You know a python-to-native compiler for windows? The closest I know is psyco and it's more like a JIT than a compiler.