DITA-OT day talks and stuff
22 views
Skip to first unread message
Ben McGinnes
Nov 15, 2015, 1:27:22 PM11/15/15
to DITA-OT Development
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Hello,
Looking at my notes/comments in the little webinar screen it appears to be more smartarse quips than not, but there are a few things that weren't (aside from my sudden EPUB 3.0 panic before the end there, which George read and also anticipated).
Regarding the comment that there aren't many involved in other open source projects also using DITA-OT, well, have you heard of a little thing called GPG (www.gnupg.org)? Yeah, that beast. Although that project uses Org-Mode for the docs and my part will also have to support reStructuredText because Pythonistas expect it (more like an API for an API because most devs don't consider compiling C headers into existing C code as an API, unlike most crypto devs; yeah, got myself a special project). Anyway, there's some old reST stuff floating around from docutils which might help with the markdown to DITA project. Docutils will already generate it's own XML from reST (which was one of the inspirations for Markdown and they behave together very well via pandoc already). So if someone were to complete an XSLT from the Docutils XML to DITA and vice-versa, that ought to solve that (probably solve quite a few things by extension, actually).
Regarding the comment that no one wants to write/dev directly in HTML. Yep, true, but you have seen what O'Reilly Media are doing with AsciiDoc and HTMLBook right? You'll either love it or cry.
A couple of points regarding the dev debate over removing all plugins by default vs. not doing so:
1) No development roadmap ever survives contact with the end users - this is the world, deal with it.
2) What you want is a plugin API so that you can predictably manage plugin "frameworks" and then let customers do what they like, but know how to change the toolkit and still be able to interact with predictable plugin behaviour types.
Regarding future breakage: you can do what you like as long as at the end of the day I can produce a fully compliant EPUB 3.0+ file with whatever in scope customisations I like and all features in the spec. While it's true I don't have the resources of IBM, well, between the crypto stuff, the politics and being a former Sun Systems Engineer I may have developed the ability to break pretty much anything ... in just the way you won't enjoy. The one bit of EPUB 3.0 breakage issues I don't expect you to do anything about is modifying existing EPUB files since XSLT can't handle setting a datetime in any timezone other than the local timezone (and the spec expects it in UTC). That issue is a minor irritation on your side of the international dateline, but since I'm 10-11 hours ahead, my side will break things. That's okay, though, I've got a Python script to rezip one of those files after correctly rewriting the timestamp via the datetime module. Said script can also run with Jython inside a JVM, so can theoretically be added to a customised DITA-OT and/or oXygenXML Editor. Although said script os a POC thing and a bit crap (ugly).
Regarding that side project thing for moving OT processing to servers/the cloud. For the corporate scenario they can still write output to a "local" directory by careful exploitation of either NFS or Active Directory for the Windows world. [You've spent so much time in your own carefully configured dev environment you forgot how much of the working world uses "hotdesking" resources.]
Regarding someone's call to arms to train future people in DITA-OT ... surely a DITA 1.3 project with all the "learning" specialisations will handle much of that and make resourcing the project an important lesson at the same time.
Let's see, that's mostly it. Save for a final note thanking Eliot once again since he explained how he could always save my precious EPUB output from anyone. There's also a note/query asking if any of the vested corporate interests would pay him to complete the D4P migration to DITA-OT 2.x. Because while he has my thanks, we'd all prefer he get their cash for stuff. :)
Regards,
Ben
-----BEGIN PGP SIGNATURE-----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=vnA9
-----END PGP SIGNATURE-----
Hash: SHA512
Hello,
Looking at my notes/comments in the little webinar screen it appears to be more smartarse quips than not, but there are a few things that weren't (aside from my sudden EPUB 3.0 panic before the end there, which George read and also anticipated).
Regarding the comment that there aren't many involved in other open source projects also using DITA-OT, well, have you heard of a little thing called GPG (www.gnupg.org)? Yeah, that beast. Although that project uses Org-Mode for the docs and my part will also have to support reStructuredText because Pythonistas expect it (more like an API for an API because most devs don't consider compiling C headers into existing C code as an API, unlike most crypto devs; yeah, got myself a special project). Anyway, there's some old reST stuff floating around from docutils which might help with the markdown to DITA project. Docutils will already generate it's own XML from reST (which was one of the inspirations for Markdown and they behave together very well via pandoc already). So if someone were to complete an XSLT from the Docutils XML to DITA and vice-versa, that ought to solve that (probably solve quite a few things by extension, actually).
Regarding the comment that no one wants to write/dev directly in HTML. Yep, true, but you have seen what O'Reilly Media are doing with AsciiDoc and HTMLBook right? You'll either love it or cry.
A couple of points regarding the dev debate over removing all plugins by default vs. not doing so:
1) No development roadmap ever survives contact with the end users - this is the world, deal with it.
2) What you want is a plugin API so that you can predictably manage plugin "frameworks" and then let customers do what they like, but know how to change the toolkit and still be able to interact with predictable plugin behaviour types.
Regarding future breakage: you can do what you like as long as at the end of the day I can produce a fully compliant EPUB 3.0+ file with whatever in scope customisations I like and all features in the spec. While it's true I don't have the resources of IBM, well, between the crypto stuff, the politics and being a former Sun Systems Engineer I may have developed the ability to break pretty much anything ... in just the way you won't enjoy. The one bit of EPUB 3.0 breakage issues I don't expect you to do anything about is modifying existing EPUB files since XSLT can't handle setting a datetime in any timezone other than the local timezone (and the spec expects it in UTC). That issue is a minor irritation on your side of the international dateline, but since I'm 10-11 hours ahead, my side will break things. That's okay, though, I've got a Python script to rezip one of those files after correctly rewriting the timestamp via the datetime module. Said script can also run with Jython inside a JVM, so can theoretically be added to a customised DITA-OT and/or oXygenXML Editor. Although said script os a POC thing and a bit crap (ugly).
Regarding that side project thing for moving OT processing to servers/the cloud. For the corporate scenario they can still write output to a "local" directory by careful exploitation of either NFS or Active Directory for the Windows world. [You've spent so much time in your own carefully configured dev environment you forgot how much of the working world uses "hotdesking" resources.]
Regarding someone's call to arms to train future people in DITA-OT ... surely a DITA 1.3 project with all the "learning" specialisations will handle much of that and make resourcing the project an important lesson at the same time.
Let's see, that's mostly it. Save for a final note thanking Eliot once again since he explained how he could always save my precious EPUB output from anyone. There's also a note/query asking if any of the vested corporate interests would pay him to complete the D4P migration to DITA-OT 2.x. Because while he has my thanks, we'd all prefer he get their cash for stuff. :)
Regards,
Ben
-----BEGIN PGP SIGNATURE-----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=vnA9
-----END PGP SIGNATURE-----
Reply all
Reply to author
Forward
0 new messages