Nationally Adopted XML eBook Format

[ahi]

Some DTDs and an example document of the XML eBook/source format used by the Hungarian Electronic Library (a very large and comprehensive eBook initiative managed by the Hungary's National Széchényi Library). 7,000+ free books (1/4th of Project Gutenberg's 28,000) from a pool of primarily Hungarian language books (which, obviously, is smaller than the pool of English language books)--both classics and contemporary works.

The format is both precise and concise, making it easier to work with (for purposes of conversion and generation) than either plaintext or (colloquial) HTML sources.

Comments in the DTD files are in Hungarian, but most of it is pretty self-explanatory. Also, the example file and the terminology defined by the DTDs are English though. I am, of course, be happy to offer additional explanation or translate anything anyone is curious about.

XML File:
http://mek.oszk.hu/03400/03407/03407xml.zip

DTDs:
http://mek.oszk.hu/mekdtd/academic/TEI-MEK-academic.dtd
http://mek.oszk.hu/mekdtd/article/TEI-MEK-article.dtd
http://mek.oszk.hu/mekdtd/drama/TEI-MEK-drama.dtd
http://mek.oszk.hu/mekdtd/mixed/TEI-MEK-mixed.dtd
http://mek.oszk.hu/mekdtd/prose/TEI-MEK-prose.dtd
http://mek.oszk.hu/mekdtd/verse/TEI-MEK-verse.dtd

- Ahi

[delphidb96]

Originally Posted by ahi

Some DTDs and an example document of the XML eBook/source format used by the Hungarian Electronic Library (a very large and comprehensive eBook initiative managed by the Hungary's National Széchényi Library). 7,000+ free books (1/4th of Project Gutenberg's 28,000) from a pool of primarily Hungarian language books (which, obviously, is smaller than the pool of English language books)--both classics and contemporary works.

The format is both precise and concise, making it easier to work with (for purposes of conversion and generation) than either plaintext or (colloquial) HTML sources.

Comments in the DTD files are in Hungarian, but most of it is pretty self-explanatory. Also, the example file and the terminology defined by the DTDs are English though. I am, of course, be happy to offer additional explanation or translate anything anyone is curious about.

XML File:
http://mek.oszk.hu/03400/03407/03407xml.zip

DTDs:
http://mek.oszk.hu/mekdtd/academic/TEI-MEK-academic.dtd
http://mek.oszk.hu/mekdtd/article/TEI-MEK-article.dtd
http://mek.oszk.hu/mekdtd/drama/TEI-MEK-drama.dtd
http://mek.oszk.hu/mekdtd/mixed/TEI-MEK-mixed.dtd
http://mek.oszk.hu/mekdtd/prose/TEI-MEK-prose.dtd
http://mek.oszk.hu/mekdtd/verse/TEI-MEK-verse.dtd

- Ahi

Kewl. Glad to have these.

Derek

[Sabardeyn]

What makes the efforts of the Hungarians in creating and using this format / schema so great? How does it improve upon any of the existing file standards?

[ahi]

What makes the efforts of the Hungarians in creating and using this format / schema so great?

I don't know if anything does. I thought the fact that this was a standard created and actively used by a National Library authority noteworthy though.

How does it improve upon any of the existing file standards?

It--mostly--marks up stuff semantically. It's an excellent source format for conversion into whatever output/display format is preferred.

Not that I know of any tools that use it as a source (though the National Széchényi Library doubtless uses such tools internally, and I do plan on writing one myself as well), but any conversion tool could more easily (and with greater confidence) generate valid output documents from this sort of an XML format than it could from either plaintext, HTML, or anything else in-between.

It's a bit like docbook, I suppose, except that it was designed by a national library and has already been (presumably sufficiently successfully) put to use to a very diverse set of 600+ documents. Put to use by a single authority, without unnecessary variation in the format or its implementation. (There's obviously lots more docbook books out there... but my experience in the past gave me the impression that in many cases it is a format that either limits the fidelity of the content or requires ad-hoc extension of it.)

I think that's the best I can answer the question... will it do?

- Ahi

[Sabardeyn]

In retrospect, I think that first question didn't quite come out the way I meant it. My apologies.

Your comments to the second question answered my poorly worded request though. Since I'm looking to future-proof my library now (before it gets any larger or harder to manage), I'm trying to find good tools and practices to implement now. Having pretty much decided on EPUB, I was confused by the sudden appearance of a new contender.

Do you have a "smallish" ebook that uses this TEI-MEK structure? (Not sure of this format's name.) I'm only looking for a couple of sample pages tied together into a book with various example implementations of the XML scheme. I don't read Hungarian and the DTDs were pretty much over my head. But I might be able to figure things out with an example file. I probably should be learning XML - but I would rather wait for improved tools that can provide assistance and automation of tasks.

Edit: Or I could just re-read your original post and download the sample book!

[ahi]

Originally Posted by Sabardeyn

In retrospect, I think that first question didn't quite come out the way I meant it. My apologies.

Do you have a "smallish" ebook that uses this TEI-MEK structure? (Not sure of this format's name.) I'm only looking for a couple of sample pages tied together into a book with various example implementations of the XML scheme. I don't read Hungarian and the DTDs were pretty much over my head. But I might be able to figure things out with an example file. I probably should be learning XML - but I would rather wait for improved tools that can provide assistance and automation of tasks.

Edit: Or I could just re-read your original post and download the sample book!

No problem.

If you go here:

http://mek.oszk.hu/katalog/index.phtml

And select in one of the dropdowns "Formátum" (i.e.: "format") and type in "XML", and click on the "keres" (i.e.: "search") button in the lower right, you will have all 600+ books listed. Although for some reason a couple of them doesn't link to their XML versions... most do though, and a few are English books.

- Ahi

[JSWolf]

But what really makes these any better then ePub?

[Nate the great]

Originally Posted by JSWolf

But what really makes these any better then ePub?

Epub isn't XML; it has XML as one of its parts. But it also has HTML tags. A strict XML file shouldn't have anything other than XML tags (I think)*.



* Could someone correct me if I'm wrong?

[ahi]

Originally Posted by JSWolf

But what really makes these any better then ePub?

You said it, I didn't.

However, since you did: The fact that its semantic mark-up instead of hodge-podge HTML aimed at getting finicky display software to cooperate makes it better... at preserving semantic and formatting information in a given text.

In fact, in another thread, I have explicitly stated that ePub is probably the most ideal format for archiving by the non-professional-Librarian, given all considerations.

Originally Posted by Nate the great

Epub isn't XML; it has XML as one of its parts. But it also has HTML tags. A strict XML file shouldn't have anything other than XML tags (I think)*.

* Could someone correct me if I'm wrong?

You're not wrong. An XML file usually is about specifying (largely, or entirely semantic) information... whereas HTML is also significantly about cajoling a myriad not 100% compatible devices and software into accidentally doing the right thing against all odds.

As a result, despite the existence of the heavenly ideal of sensibly formatted and standard-compliant HTML markup, the average document is going to be somewhat to moderately messy and difficult or (over multiple documents created by different authors) impossible to automatically parse/convert in a definitively correct way (unlike a well defined semantic XML format like this).

- Ahi

Ps.: Any suggestions, JSWolf, on how to improve my ugly duckling of an ePub file of The Art of War (posted in another thread)? I find the lack of interest therein--particularly in light of the rather greater positive attention the PDFs received--to be odd given the apparent enthusiasm about ePub on this forum. Particularly since I spent more time making the single (and, on my Sony PRS-505, not particularly great looking) ePub, than I did producing the 8 considerably more professional looking custom PDFs.

[Peter Sorotokin]

Originally Posted by Nate the great

Epub isn't XML; it has XML as one of its parts. But it also has HTML tags. A strict XML file shouldn't have anything other than XML tags (I think)*.

* Could someone correct me if I'm wrong?

You are somewhat wrong. EPUB has XHTML tags (and syntax), not HTML ones. And XHTML is XML (unlike HTML which is SGML).

XHTML attempts to be a semantic language. XHTML has well-defined DTD and can be validated. Most HTML presentational attributes and tags are either outlawed (e.g. FONT) or deprecated (ALIGN). Instead, CSS stylesheets do the job. In addition, classes can be assigned to add your own semantic structure on top of what XHTML provides.

In reality, a lot of XHTML content is just somewhat tidied-up presentational HTML - and that looks really ugly on the source level. It is still XML, but there is not much sematics in it, it is just styled to look right. I consider this to be bad authoring practice, but that's somewhat a question of opinion.

There is no such thing as "XML tags". Better to say, there are no tags defined by XML specification itself. XML defines how to build various XML dialects - and XML dialect is where tags are defined. XHTML is merely one of such dialects that tries to make transition from HTML simple. (SVG and DTBook, also mandated by EPUB spec, are also XML dialects).

[ahi]

You are right, of course, Peter. I generally don't bother to make the finer distinction, but arguably XML is a meta-language for creating XHTML-like languages.

My point above was that while HTML/XHTML must yield semantic coherence and/or consistency to some degree in favour of actually having it displayed right across the board, an XML language that is not concerned with display logistics is more free to mark information up more concisely and succinctly, and eminently parseably (with the program less likely to get confused what a given combination of tags actually mean).

Am I making sense? (Despite, I suspect, having become somewhat less understandable for the less tech-savvy.)

- Ahi

[Jellby]

Originally Posted by ahi

My point above was that while HTML/XHTML must yield semantic coherence and/or consistency to some degree in favour of actually having it displayed right across the board, an XML language that is not concerned with display logistics is more free to mark information up more concisely and succinctly, and eminently parseably (with the program less likely to get confused what a given combination of tags actually mean).

Perhaps the following can be an illustration.

I just realized that XHTML forbids block content inside a <p> element. So you cannot have a <div> inside a <p>. Is this sensible? Maybe, but I wanted to have <div>'s inside <p>'s for cases when a piece of poetry is inserted in a character's speech, and the text resumes after it with no indentation or whatever. I can work around this by having a <p>, then a <div> and then a <p> with no indentation, but the logical structure would be a <div> inside a <p>.

By not being limited by XHTML, this issue can be avoided. Is this the kind of thing you mean?

[kovidgoyal]

@Jellby: The correct way to do poetry in XHTML would be

<div class="poem">
<div class="stanza">
<div class="line">
...
</div>
.
.
.
</div>
.
.
.
</div>

Set the text-indent property in CSS to have the div mimic the behavior of a <p>

[ahi]

Originally Posted by Jellby

Perhaps the following can be an illustration.

I just realized that XHTML forbids block content inside a <p> element. So you cannot have a <div> inside a <p>. Is this sensible? Maybe, but I wanted to have <div>'s inside <p>'s for cases when a piece of poetry is inserted in a character's speech, and the text resumes after it with no indentation or whatever. I can work around this by having a <p>, then a <div> and then a <p> with no indentation, but the logical structure would be a <div> inside a <p>.

By not being limited by XHTML, this issue can be avoided. Is this the kind of thing you mean?

This is what the TEI-MEK XML format uses for poetry (and its not embedded in a paragraph, but presumably could be for the very reason you mention):

 <lg type="verse">
  <l>Mivel malommester alsóknak hadnagyok,</l>
  <l>Szárnyas madaraknak az sólyom az urok,</l>
  <l>A fenevadaknak oroszlány királyok,</l>
  <l>Az fáknak fenyõfa légyen az császárok.<ref target="note5221-1" type="footnote">[<hi vertical-align="top">96</hi>]</ref><note id="note5221-1" type="footnote">96. RMKT IV (Budapest, 1967), p. 439.</note></l>
 </lg>

My point is less about not being limited by XHTML, and more that XHTML has no definitive way of encoding inserted poetry, whereas an XML dialect (like the TEI-MEK one) that aims at accurately and concisely encoding semantic information, as you see, does.

A program written to parse the this XML format encountering the above code should be unambiguously clear as to how to interpret it. On the other hand, your intuitive way of encoding poetry in XHTML is certain not to precisely match the way a 100 other random people do so. Assuredly there would be overlap... but surely there would easily be a dozen or two different ways, all nominally correct, but none definitively correct and certain to be used by all "correct" documents.

Does that clarify my meaning?

- Ahi

[Nate the great]

Originally Posted by Peter Sorotokin

You are somewhat wrong. EPUB has XHTML tags (and syntax), not HTML ones. And XHTML is XML (unlike HTML which is SGML).

XHTML attempts to be a semantic language. XHTML has well-defined DTD and can be validated. Most HTML presentational attributes and tags are either outlawed (e.g. FONT) or deprecated (ALIGN). Instead, CSS stylesheets do the job. In addition, classes can be assigned to add your own semantic structure on top of what XHTML provides.

In reality, a lot of XHTML content is just somewhat tidied-up presentational HTML - and that looks really ugly on the source level. It is still XML, but there is not much sematics in it, it is just styled to look right. I consider this to be bad authoring practice, but that's somewhat a question of opinion.

There is no such thing as "XML tags". Better to say, there are no tags defined by XML specification itself. XML defines how to build various XML dialects - and XML dialect is where tags are defined. XHTML is merely one of such dialects that tries to make transition from HTML simple. (SVG and DTBook, also mandated by EPUB spec, are also XML dialects).

Thank you for clearing that up. While I've used XML for some time, I only just started teaching myself the fundamentals.