Just had a short look at the documentation of SwordHammer and think this is a promising concept. Specifically for some very complicated texts delivered to us as MS word etc files. I will most certainly have a go at using it.
Peter
Sent from my mobile. Please forgive shortness, typos and weird autocorrects.
-------- Original Message -------- Subject: [sword-devel] Introducing myself, SwordHammer, and asking a ton of questions From: Tom Sullivan To: sword-devel@crosswire.org CC:
Y’all:
This will be a long one, so let me first summarize, then provide details for each section:
Personal Introduction: My name is Tom Sullivan, i...@beforgiven.info.
Context of my interest in this forum: I have developed a program, SwordHammer, to translate from WYSIWYG word processor output to OSIS and Sword Modules. It is in beta, and is at beforgiven.info/SwordHammer. I hope it will be useful to some of you, and to a wider audience of authors, editors, and translators.
Questions: But, there are issues with my understanding of OSIS, Sword, and what is to be expected of front end programs, and of the back ends, jsword and sword. I have read the docs fully and more than once, but there are places where there are contradictions, things that are unclear to me, and undoubtedly, things that I stupidly missed. So I would appreciate some help. I need this information so I can make SwordHammer do its job well. I have extensively tested it; it mostly works well, but there are warts and blemishes. A prime goal is to isolate the user from technical details.
Perhaps I could help with the documentation? It seems to need work. Perhaps Crosswire could specify a simpler OSIS if nobody else is using it?
Future Direction: Someday I hope to write a front-end for interlinear Bible presentation. A very simple front end sync standard would be helpful.
More details:
*** Personal Introduction:
My name is Tom Sullivan and you can reach me at in...@beforgiven.info. I am a retired engineer, but in most jobs was also IT administrator or involved with it. I started programming in FORTRAN on the IBM 360 mainframe in 1972. About a year and a half later, by God’s sovereign grace alone, the Holy Spirit regenerated my heart and gave me faith in Jesus Christ whose bondslave I have been ever since (not that I have always acted the part very well).
What follows about myself is not all that important, but may perhaps help others understand my strengths, weaknesses, and motivation.
At the end of my career I was mostly using C, C#, and (legacy) VB6. When Microsoft released that unnatural chimera called Windows 8, started putting spyware (“telemetry”) in Windows 7, tried to cram Windows 10 (replete with spyware and ads) down my throat, and etc., a divorce was in order. (Microsoft may possibly have a somewhat different perspective.) I settled on Debian and tried to stay with C# using Mono, but it was buggy, and then Microsoft bought Xamarin (who owned Mono). Forced to choose a new language, I learned Python 3 and have written an awful lot of Python code to convert my own stuff over from Windows to Linux. Although I must still use Windows for a few unmovable commercial programs, I am now almost fully on Debian, still learning, and still considering myself to be a Linux newbie even after two years.
Of course, Linux is a different world. The quantity and quality of the software is amazing, especially considering its largely volunteer origin, and the price is right. But everywhere, documentation tends to be a weak spot. What struck me most is that, compared with commercial offerings, there is a lack of good Bible programs for Linux. The relative lack of modern, translations, commentaries, and reference material is even worse. Those of us who are Linux users are, in effect, an unreached people group. It is clear that there are many who are hard at work on Bible programs and that they are making good progress to remedy the gap. With regard to input material, this obvious need is what has motivated me to tackle the project of writing SwordHammer in addition to my other retirement job of translating Puritan and other classic works into modern English. (Of course, I hope to use SwordHammer to publish my own works as well.) Readers who have used programs like Libronix will understand the great value of linking Christian works and their Bible references to the actual Bible passages. Too many readers do not bother to look up references. This is in addition to linking commentaries, dictionaries, etc., to the Word of God to facilitate careful study.
So OK, I am the new employee who has still to find the bathrooms. I have a lot to learn. But perhaps I can also bring some helpful outsider’s viewpoints and suggestions. I have waited until now to join this list because, first, I wanted to learn and test my learning about OSIS and Sword modules, and second, because I wanted to present SwordHammer in beta as an indication that I am serious about helping out with the Sword project and making the whole “ecosystem” better, if God will so allow me.
*** Context of my interest in this forum:
I have developed a program, SwordHammer, to translate from WYSIWYG word processor output to OSIS and Sword Modules. It is in beta, and is at beforgiven.info/SwordHammer. I hope it will be useful to some of you, and to a wider audience of authors, editors, publishers, and translators. My target audience includes the kind of people who, while they can do their jobs well, even excellently, on their computer, some of them will need to call for support for almost every IT problem. I have written the instruction manual accordingly.
I encourage you all to at least download the documentation, if not try SwordHammer as well. It seems a lot to ask, but if any of you could be so kind as to look at my documentation and point out to me places where I am inaccurate with respect to OSIS or Sword modules; such are most likely to show up in what I have termed “General Questions.”
SwordHammer uses as input the .ODT files produced by LibreOffice, which (should) conform to OASIS, which is an international standard, ISO/IEC 26300-1.
The essential working principle of a WYSIWYG word processor is to enable a user to present a document to a reader in terms of how the document looks to the reader; the word processor is largely unaware of how the user derives meaning or information from the appearance of the document.
By contrast, SwordHammer’s job is to extract meanings from appearance and translate those meanings into OSIS, then Sword Modules. SwordHammer does this by analyzing the document and picking out all of the section, paragraph, font, and other information pertaining to the text. We will call these kind of things “attributes.” Each time SwordHammer encounters a new set of attributes (even if only one attribute differs) it generates a question for the user as to its meaning (verse, footnote, italic, title, and so on). The questions are presented to the user in a (new) question document that contains the user’s original input with questions interspersed in context. When the user answers those questions, SwordHammer learns the meaning of the formatting and is hence able to translate the document into OSIS. If one thinks about it, this is not an unreasonable approach; readers distinguish between items in a document by appearance also.
SwordHammer also asks a large number of questions which are driven in part by the requirements of OSIS and Sword Module generation. So far, SwordHammer has been tested on Bibles, Commentaries, and Generic (General) Books.
It is my belief that SwordHammer is greatly easier to use, more productive, and less error-prone than forcing authors to learn OSIS and about Sword Modules. SwordHammer attempts to insulate and isolate the author or document editor, etc., from having to learn and interact with either OSIS or the details of Sword Modules. (SwordHammer currently does not support all OSIS features, and even fully developed may never get there, but I hope that it is a good start.)
In developing SwordHammer I have, of course, interacted deeply with Sword Module creation and OSIS. There are a number of issues and questions that have arisen that will appear below. Obviously, since SwordHammer is an automated system, I will have a bias towards using OSIS in the most consistent and simple manner possible. This bias may perhaps show in some of my questions.
*** Questions (* marks a new question)
Before going further, let me just say that I fully appreciate the fact that software development is hard work, that the work is being done by unpaid volunteers, that it is painful, but necessary when someone tells you that your program has a bug, and that not all back ends or front ends will have the same feature set. My goal here is to learn so that I may develop SwordHammer to its fullest potential as the Lord gives me health, strength, and opportunity.
In particular, I have been testing with these front ends: Xiphos 4.04 , BibleTime 2.10.1, and BibleDesktop 2.0 Beta. My reports of problems with these is not to slam them, but they just happen to be three that work on my Debian Stretch system. These may not be up to date. For example, I have been informed by Xiphos that version 4.0.7 is current. But here is why I use what may be old versions: I am expecting some of my users to be unsophisticated; to expect them to compile from source is unrealistic. It also seems unrealistic to expect them to use other than stable distros. Debian does support package updates for bug fixes; may I please ask developers to kindly supply such fixes as needed.
* When first studying the documentation for OSIS, I looked in vain for further information. I was finally able to make contact with SIL, and they said that they no longer used OSIS, even though they were involved with its development. My impression is that only Crosswire supports OSIS anymore. * Is this correct? * If it is correct, should the OSIS specification be pared down for simplicity? For example, does it really matter if a section is tagged “colophon” or “gazetteer” in terms of how front ends will handle the material? Similarly, there are a number of what may be seen as subdivisions of the general category of “translator’s notes,” such as “transChange,” “translation,” or “variant.” Will not front ends present all of these as non-cross-reference notes in the same way? Which, if any, do actually deserve special treatment? (I realize this is a general question and only expect general answers at this time.) In particular, seems to me to be completely useless (except for its use for red-letter). Why not just pass through the source material’s mark for the language being used?
* Would it be worthwhile to re-write the OSIS documentation to reflect the needs of Crosswire and incorporate the articles in the Wiki into the same document, so there is a comprehensive and consistent guide? And, yes, being an experienced author and editor, I believe I would be willing to volunteer for this job if I can get both good input and feedback; right now I don’t know enough just yet.
*** I present the following bug reports. Implied in all of them are the following questions: * Is this something that should work, but I must be doing something wrong? If so what? * Is this something that will never be supported? * Is this something that needs to be fixed? * Etc. I am looking for education here. Many thanks.
* Using: osisCore.2.1.1-cw-latest.xsd, there are Work sub elements that give errors:
* is not accepted in chapter title by osisCore.2.1.1-cw-latest.xsd
* In both Xiphos and BibleTime: . Tabs not recognized . and delineated poetry does not work reliably . Fonts are random sizes (sometimes) . Images do not display description on mouseover for vision impaired . underline does not work . small caps does not work . strikethrough does not work . tables do not recognize elements . fail to properly handle 0x2019 right single quotation mark in a title when used as an apostrophe. A title is truncated. In regular text, it does not appear. Note: 8217 2019 ’ RIGHT SINGLE QUOTATION MARK is the recognized standard for the English apostrophe. . Non-ASCII characters do not print, in spite of documentation that says to use UTF-8. BibleDesktop does not have this problem. . fail with words of Christ in red if who="Jesus" used in quote milestone (as opposed to non-milestone). The red never shuts off.
* BibleDesktop does not display all notes if there are a lot of them.
* On pg 67, the OSIS manual shows to make second level list items. However, this produces extra bullet points. The correct manner is to just use , where is a child of the previous list. Bibletime and Xiphos then display correctly, but BibleDesktop shows an extra bullet point. ALL programs insert their own markers in spite of the user supplying their own, such as numbering. HOWEVER, this causes OSIS validation errors, AND if done the way of the OSIS manual, you do not get validation errors, but you do get double bullet points for nested entries.
* In Windows, xml2gbs that is loaded by Xiphos produces only a .bdt file. The same XML file works fine for xml2gbs Linux version. There are no error messages generated. The output is similar, but while the Linux version gives titles, the Windows version gives the value of the osisID. The problem is the version that comes in Xiphos. The downloadable from http://www.crosswire.org/ftpmirror/pub/sword/utils/win32/ works, but the file versions seem old.
* In contrast to documentation, VC2012 redistributable is needed.
* A footnote marker of a dagger or special character will cause Xiphos and Bibletime to mishandle the footnote. BibleDesktop works fine.
* BibleTime does not display an image in a generic book.
* BibleTime ignores paragraphing.
* I have not figured out how to include non-canonical and separate introductory or extended explanatory text in a Bible. ESV2011 does it however. How do you do that?
_______________________________________________
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
