2005

January

1^st January 2005  ⁴⁰¹  Well Wishes for the New Year. To quote one of the most generous New Year wishes received, wishing you (all) "a prosperous 2005 full of happiness, health, joy, and success" and thanks for that.

3^rd January 2005 Guess this should have been looked at earlier, added:

Free as in Freedom: Richard Stallman's Crusade for Free Software, by Sam Williams

"Gnu Public License 2", (GPL 2) Free Software Foundation

4^th January 2005 SiSU release announced.  ⁴¹² 

Vertigo, by Leena Krohn

There have already been many subsequent updates - for the current release details see:

11^th January 2005 A more complete bundle has been posted.

27^th January 2005 SiSU separate Stable and Development Branches are needed and will be introduced shortly, based on a version numbering policy.  ⁴¹⁷ 

31^st January 2005 Apple Mac, first look, very little problem getting SiSU to run, html, ascii, xml generated. There will be issues to resolve, but not major ones:  ⁴¹⁸ 

sisu -v

SiSU Sabaki: version 0.1.2-19 of 2005w05/1 (20050131) & Ruby: ruby 1.8.2 (2004-12-25) [powerpc-darwin7.7.0]

my info:

SiSU Sabaki: version 0.1.2-19 of 2005w05/1 (20050131) & Ruby: ruby 1.8.2 (2005-01-10) [i386-linux]

February

3^rd February 2005 SiSU installation Rakefile extended to do post ordinary install/setup and try generate first document. Release 0.1.2-23 worked without hitch on testers Debian box, producing html, ascii, html, xml and pdfs ... that release is withdrawn as it does a bit too much in root. Subsequent release, only now undergoing tests, will attempt to do the same after requesting ordinary user privileges if sudo is used.

4^th February 2005 SiSU installation starting to work more smoothly, tested on Debian and Fedora using Rake. Thanks G.A.

5^th February 2005 Great email received to inform me that Tainaron   ⁴¹⁹  by Leena Krohn   ⁴²⁰  has "appeared in print in the USA"   ⁴²¹  and SiSU played a small role in this. Apparently after it's publication on the web, it got postive reviews,   ⁴²²  and so was "discovered" and sought for publication..., "it found its way to being printed in English via your web edition"  ⁴²³  Tainaron continues to have good reception.  ⁴²⁴  The printed edition has images that were later included also in the online edition. Amazon link here,   ⁴²⁵  if you wish to order this handsome book, (see other links provided within this note if you wish to read an online edition). [I received an autographed copy of the US printed edition from the author on March 4^th, 2005 :-) ]

12^th February 2005 Stable and development branches introduced, the 0.1.4 and any subsequent even number endings are (relatively) stable, and the 0.1.5 (and subsequent odd number endings) denote the development branch. Changes made in the development branch will periodically be rolled back into the stable branch as found to be appropriate.

0.1.5 development branch opened, 0.1.4 branch will be "stable"

SiSU Sabaki 0.1.5-0

15^th February 2005 SiSU Sabaki 0.1.4-0 (0.1.4 is stable branch)

16^th February 2005 Confirmation today that basic SiSU install with Rake has no problems on Gentoo, thanks G.A. Current list Debian, Fedora, Gentoo.

18^th February 2005 Google Group created SiSU e-docs   ⁴²⁶  to be very low volume, and moderated at least to start with. Mostly experimenting with use of "Google Groups beta".

21^th February 2005 Placed SiSU development branch in directory, which may be browsed or downloaded  ⁴²⁷  (instead of making 3Mb + tarballs each time):

The University of Oslo online resource though generous, offers a limited feature set.

March

2^nd March 2005 Some work done in development branch on internationalisation.

Default locale is to be Unicode UTF-8 which LaTeX, Postgresql (and Vim) can use.  ⁴²⁸  ...  ⁴²⁹ 

Stubs for easy changing of document size from the default A4 in LaTeX are in place,  ⁴³⁰  though the text dimensions and placement for the alternative default sizes text need further refinement.

6^th March 2005 Cosmetic change to default html, css  ⁴³¹  and change to default navigation arrow buttons.  ⁴³² 

9^th - 12^th March 2005 SiSU Sabaki 0.1.4-1   ⁴³³  followed closely by SiSU Sabaki 0.1.4-2   ⁴³⁴ 

Added first entry   ⁴³⁵  to the RAA, Ruby Application Archive   ⁴³⁶ 

SiSU

A command line driven (batch processing) document generator with: minimal (simple & mnemonic) markup requirements (which include the possibility of providing Dublin core or other semantic information about each document); multiple output options (file creation & database populating: including plaintext, html, xhtml, XML, LaTeX, pdf, and populating of PostgreSQL relational database) all outputs sharing a common citation numbering system.

The database loading feature is not required but provides additional possibilities, for example for database searches the database is populated such that database match results are available at a text object (roughly paragraph level) either as an list of documents together with: linked object citation numbers for all text objects (paragraphs) matched; or displayed text for each text object (paragraph) matched.

markup in your favourite editor, (vim syntax highlighting file provided)

share output off: filesystem, web, database

SiSU is not for technical documentation.

Summary

‹http://www.jus.uio.no/sisu/SiSU/1.html#summary›

Book Samples

‹http://www.jus.uio.no/sisu/SiSU/examples.html#books›

Markup Examples:

‹http://www.jus.uio.no/sisu/SiSU/examples.html#markup›

‹http://www.jus.uio.no/sisu/sample/›

Markup Syntax summary:

‹http://www.jus.uio.no/sisu/sisu_markup_table/markup.html›

SiSU development dates back many years. (documents need rewriting, they predate the release of SiSU).

man pages, interactive help provided.

10^th March 2005 At last a good book on zsh: "From Bash to Z Shell" by Oliver Kiddle, Jerry Peek, and Peter Stephenson; isbn: 1-59059-376-6

14^th March 2005 Look again seriously at CSS.

‹http://www.csszengarden.com/› as a demonstration of what can be done with CSS is impressive. This page has been put forward by others on a number of previous occasions, but I finally got round to looking at it carefully enough to appreciate what it is (after the most recent posting/comment referring to it on ruby-lang).

27^th March 2005 Finally persuaded myself to permit the building of the source document in parts, i.e. with requests from a master document, or more generally to permit a document to be processed to make requests for information from another file.

The documents to be incorporated may be complete sisu markup documents, suitable for independent processing by sisu (suffix .s1 .s2 or .s3 [Naming of file-extensions changed 200603]) or they may be fragments of sisu markup syntax (suffix .si).

Three types of document require statements are permitted, ordinary sisu syntax, textlink style syntax, or thlnk style syntax (the latter two permit hypertext ascii calling of the constituent parts and editing in the vim editor).  ⁴³⁷  For information on how this is implemented, or rather to be used, see the documentation (currently development branch version manifest).

There has been pre-processing, actually a more sophisticated merging of documents standard forms against templates (or termsheets) since it was developed in the legal domain since 1998, but this is a nice feature to have as well.

This document is a prime candidate for being split into parts. It is long, and the sections from which it is constituted are fairly independent (and could stand on their own), apart from this it is due for an overhaul having much that is off-topic, and much that is better presented elsewhere with respect to SiSU (e.g. in the man pages, or in markup text files rather than in images). An explanation for this being the case, is that earlier versions of this document were placed on-line several years before the code for SiSU was released.

This document has been split into several parts.

Also, urls may be provided instead of filenames for processing remotely located sisu markup.

April

10^th April 2005 Stable and development branches version number changes. The stable branch is now 0.2 and the development branch 0.3.

Will use a three digit scale to track changes instead of four (0.0.0 instead of 0.0.0-0).

Changes made in the development branch will periodically be rolled back into the stable branch as found to be appropriate.

There have been quite a few updates in recent releases, tracked in the version manifest and the section on downloads (current and previous versions).

12^th April 2005 Stable and development branches version number changes. The stable branch is now 0.4 and the development branch 0.5.

Changed representation of image dimensions, without retaining backward compatibility, a rare event that has the potential to cause breakage... second digit version number change is for this reason.

13^th April 2005 Re-introduce sqlite module.

16^th April 2005 Debian Sid includes ruby1.9, first look, for SiSU, that's how lazy Debian keeps me, may Debian keep it that way.

19^th April 2005 New Branch Started, 0.7. Removed old markup style from SiSU, which has lived alongside the current syntax for a while, as it adds nothing and it is easier to describe SiSU markup without.  ⁴³⁸ 

24^th April 2005 New branch 0.6 ready.

26^th April 2005 I am beaten back to an old idea, wrestled with at the time of introducing ocn, that is, to have alternative forms of ocn object citation numbering. In addition to what is provided now,  ⁴³⁹  distinguishing headings from other objects/paragraphs.

I end up for the time being with three parallel numberings, provided in the meta markup stage of each document and available for subsequent processing: *i* regular ocn; *ii* ocn with headings separated from other objects; and *iii* ocn with each different type of object numbered and identified according to type of object (heading, paragraph, table, image, etc.). It is still possible to omit regular numbering, in which case the object belongs to the un-numbered class and is numbered accordingly.

The need comes about because of a significant class of documents that have paragraph numbering, assigned by author/publisher, that is very similar to the ocn automatically provided by SiSU, and that in some cases it is necessary to match (because the similarity brings about confusion). This is achieved more "automatically" by having such options. In many cases headings are not give object numbers, and the author does not number every paragraph separately - such a document is more easily matched using a combination of the second numbering provided, (together with manual markers to escape un-numbered paragraphs).

Other (optional) candidates for future inclusion in the ocn capsule are: *a.* where there is an official publication, manually included page number markings of the official at the start of each page, these would then be automatically added to each object that occurs on each page (i.e. each objects that occurs on each page of a book (e.g. page 37) would carry that information); *b* heading number detail, these are currently extracted and made available, (e.g. to the database module) by other means.

28^th April 2005 New branch 0.8

May

3^rd May 2005 At last Debian packages for SiSU. They need to be downloaded manually though, from: ‹http://www.jus.uio.no/sisu/debian›

Version released sisu_0.10.0_all.deb and sisu-examples_0.10.0_all.deb

5^th May 2005 Started using Darcs-buildpackage   ⁴⁴⁰  together with Rant   ⁴⁴¹  for building sisu debian packages. Darcs   ⁴⁴²  in light use for over a year.

6^th May 2005 An early Debian package of SiSU sisu_0.10.1-1_all.deb and sisu-examples_0.10.1-1_all.deb

8^th May 2005 Early Debian package of SiSU, paths sorted:  ⁴⁴³ 

which maps to:

17^th May 2005 Headway with Debian packaging, much automation. Still detail to sort, and have not checked all dependencies.

20^th May 2005 Debian apt-get repository made for sisu.

Apt sources list should be set to:

deb ‹http://www.jus.uio.no/sisu/pkg› ./deb/

information on the source should be available from:

‹http://www.jus.uio.no/sisu/sisu_download/download.html#current›

‹http://www.jus.uio.no/sisu/SiSU/download.html#current›

Unification of packaging efforts, results in switch from using sha to md5 digests to identify tarball contents. The same tarball as is released is the debian original.

21^th May 2005 Changed package directory directory structure, and references to it.

June

17^th June 2005 On the anniversary of my meeting with IBM,   ⁴⁴⁹  an interesting event, which led to a chain of other interesting events, I post a copy of my abandoned U.S. provisional patent application   ⁴⁵⁰ 

Of particular interest is the ease of streaming documents to a relational database, at an object (roughly paragraph) level and the potential for increased precision in the presentation of matches that results thereby. The ability to serialise html, latex, xml, sql, (whatever) is also inherent in / incidental to the design.

My reasons for abandoning it may be summarised as follows:

1. I have serious reservations about software patents, and the conflict between their claimed intended purpose (encourage invention) and what they actually do, or are used to do (hinder, slow down innovation).

a. After having used Gnu/Linux for several years, (and this experience in relation to software patents), my conclusion on Software Patents and their effect on software innovation, software quality and reliability, is that they are problematic and do more to hinder than to promote innovation.

b. More software innovation and software diversity occurs in the space that exists without them, and the software that results is frequently more reliable and of higher quality. The GPL ultimately secures this space for software.

2. I also had problems with my starting the patent process so late in the development of SiSU. Had I intended to go this route, I should have started the process several years ago.

a. SiSU is based on ideas that date back to my time at the University of Tromsø.  ⁴⁵¹  Indeed with the basic idea, I went about equipping myself with the best tools I could find to implement it.  ⁴⁵² 

b. The question of publication was in issue. I was always open about what I was trying to do: I gave presentations of SiSU from the start (though it had not yet been named) to whoever was interested, and at a couple of convenentions; I published the first versions of this web page on the subject in November 2002 describing SiSU in considerable detail, and; prior to that there were notes on the website lexmercatoria,  ⁴⁵³  and; the output of lexmercatoria gave some indication of what SiSU was trying to achieve. I took the ultimate step in publishing SiSU _sauce_ on January 4^th, 2005 when the source code for SiSU was placed to the Net and announced on ruby-talk.

3. The cost of acquiring a Patent is high (tens of thousands of US dollars),  ⁴⁵⁴  and the cost of defending a Patent, should it be violated is prohibitive (hundreds of thousands of US dollars). There is a high threshold of entry into the Patent game. Patents are primarily for large players, or those that can afford to play, and/or wish to sell to larger players.

Why then did I start the process in the first place? A good question. The short answer would be, to give myself time to think.  ⁴⁵⁵ 

A longer answer. After the meeting with the IBM software innovations evaluator,  ⁴⁵⁶  it was indicated that this idea could be of value to them.  ⁴⁵⁷  He also indicated that to be prepared to pay anything substantial for this technology coming from me, there would have to be a patent, otherwise they would prefer to prepare their in-house equivalent, and they would not mind employing the number of coders it took for the time it took, he explained that this was their way. This led to interesting discussions on the subject of Software Patents. He thought they might actually prefer to (have to) pay a substantial amount for such a Patent, because it gave control over the technology, i.e. they could prevent competitors from doing the same. The choice for me it seemed was between much or little, and I needed time to think. The provisional patent approach cost least and gave me a year within which to file for the Patent... but I had all the reservations about Software Patents mentioned previously.

Taking the first step on the road to procuring a Patent, resulted in some interesting thinking and correspondence. One idea was an "FSF Patent" Trust

The idea was to acquire a Patent and to grant free use of it, with the scope of free use defined as use compatible with the GPL   ⁴⁵⁸  so for all Free Software Foundation   ⁴⁵⁹  approved software licenses, and "guaranteed" by some form of trust that was held by an independent third party institution. The beneficiary in the event that the software was licensed for non-free use would be the inventor. It remains of academic (intellectual and legal) interest, whether and how such a device could have been made to work.

The potential offered by the ideas behind SiSU are interesting. The Patent would have gone for the jugular, as patents seek to do, covering a wide range of potential application, based at heart on a simple but essential idea.

SiSU is available for download and use on the Gnu/Linux and Unix operating systems under the terms of the GPL   ⁴⁶⁰  (with the possibility of an alternative license for developers unable to use the GPL. (this dual license, being what I would have sought to be able to control more effectively through a Patent held in trust).

19^th June 2005

Markup syntax is currently under review.  ⁴⁶¹ 

WARNING changes break current markup, a theme is the reduction of characters, and the removal of most unclosed curly braces where used previously. Continued goal of simplicity and minimalism, however, syntax higlighting becomes even more crucial as some important markup becomes more difficult to spot, headings in particular.

This has been implemented in the development version but testing is required. Other changes are likely.

This has been implemented in the 0.15 development branch. Extensive testing required.

21^th June 2005 Placed Cleaner Markup in stable branch as 0.16.0 no doubt prematurely as breaks old markup and much further testing is required.

Apologies for the inconvenience caused.

A vim regular expression substitution summary of the changes are provided with the program in the version_manifest.changelog and in the debian changelog.

22^th June 2005 Finally felt obliged to add SiSU   ⁴⁶²  to Freshmeat   ⁴⁶³  as a placeholder for the name.

SiSU (Structured information, Serialized Units)  ⁴⁶⁴  is is a document creation and management framework. It manages electronic texts and document collections with minimal markup requirements, and provides an "atomic search" and text positioning system. It can produce output as plain-text, HTML, XHTML, XML, LaTeX, PDF, or SQL (for PostgreSQL or SQLite). SiSU is a commandline tool/batch processor and a markup syntax. Documents are prepared by marking them up for SiSU with your favorite text editor and then generated in the desired output formats. A vim syntax highlighting file for SiSU syntax is provided.

Homepage: ‹http://www.jus.uio.no/sisu/›

Tar/GZ: ‹http://www.jus.uio.no/sisu/SiSU/download.html#current›

Changelog: ‹http://www.jus.uio.no/sisu/sisu_changelog›

Debian package: ‹http://www.jus.uio.no/sisu/SiSU/download.html#debian›

Trove categories:

[License] OSI Approved :: GNU General Public License (GPL)

[Operating System] POSIX :: Gnu/Linux

[Programming Language]: Ruby

[Topic] Database, Information Management, Information Management :: Document Repositories, Text Processing, Text Processing :: General, Text Processing :: Markup, Text Processing :: Markup :: HTML/XHTML, Text Processing :: Markup :: TeX/LaTeX, Text Processing :: Markup :: XML

Dependencies: [change]

Ruby (required)

PostgreSQL (recommended)

SQLite (recommended)

teTeX (recommended)

tidy (recommended)

Vim (optional)

July

9 - 17^th July 2005 Debconf5, Helsinki, Finland

Much interesting developments within Debian, and a great social event held in weather that was if anything a bit too sunny and hot. Debconf5 group photograph   ⁴⁶⁵  and and photos taken with my camera   ⁴⁶⁶  kindly placed online by Don Armstrong.

Several ideas related to SiSU... noted largely to remind myself, not sure what if anything will be implemented, but the framework on which such additions could be done is there and healthy.

SiSU  ⁴⁸³  was uploaded to Debian queue new from which it made its way to unstable and testing by Gunnar Wolf   ⁴⁸⁴  during debconf5, thank you.

Description: electronic document collection generator, search enabling A command line driven document generator with minimal simple and mnemonic markup (which includes the possibility of providing Dublin Core or other semantic information about each document), that produces multiple output types (including plaintext, html, XML, (OpenDocument), LaTeX, pdf, and populating of postgresql or sqlite relational database) all sharing a common citation numbering system.

The relational database feature is not required but provides interesting possibilities, including the possibility of atomic search of documents for matching units of text, primarily paragraphs that can be displayed or identified by object citation number, from which an index of matched documents and each matched paragraph within them can be displayed. _1 man pages, interactive help (and vim folds and syntax highlighting) provided.

The dummy package sisu-complete together with sisu-examples installs the whole of SiSU

Home page ‹http://www.jus.uio.no/sisu›

Also thanks to Joachim Breitner,   ⁴⁸⁵  Kaj Tuunanen, Wookey, Johannes Jordens and to: Don Armstrong,   ⁴⁸⁶  Eric Evans, Fumitoshi Ukai,   ⁴⁸⁷  Hector Oron Martinez, Matti Korttila, Paul Sladen, Mark Shuttleworth,   ⁴⁸⁸  David Moreno Garza,   ⁴⁸⁹  Hanna Heen, Philip Hands,   ⁴⁹⁰  Martin Krafft and Andreas Schuldei   ⁴⁹¹ 

18 - 2^nd August 2005 Kimito, Finland, mostly offline

14^th July 2005 Re-opened the development/experimental branch and added two md5 digests/sums/hashes to each paragraph created in the "metaverse" (machine text representation), one for content stripped of markup, the other a stamp of the prepared state in the "metaverse". They are available for use in all downstream processing. The clean text (markup-free) hash should eventually match what is streamed into the clean text field of the database, whilst the hash that includes markup will be a stamp of the metaverse prepared object/paragraph. Situation is fluid.

0. md5 digest taken of the document source and placed in the metaverse (implemented much earlier)

In addition produce md5 digests (in metaverse) for every paragraph and endnote, specifically an md5 digest for:

1. clean/stripped text without any markup, paragraph, headings etc. without endnotes

2. endnotes clean/stripped text digest only (there may be several endnotes within a paragraph)

3. whole object, text with markup and any endnotes

[digests should not include other digests]

Digests are available for use in downstream modules, e.g. in database mapping, to identical paragraphs, possibly in html scroll, to identify unchanged content, etc.

Implemented 20^th July 2005, but need to be consistent about content of objects on which md5 digest is made, so relates directly to sql clean.

19^th July 2005 SiSU   ⁴⁹²  accepted in Debian Sid, it was uploaded at the end of debconf by Gunnar Wolf.

SiSU thanks Ruby, Gnu, Linux & Debian

22^nd July 2005 0.18 branch (development branch 0.17 rolled back in to main), implementing md5 digests on objects, see earlier comments. Structural changes in place. Work on consistency of content of objects on which md5 digest is made, so relates directly to sql clean.

24^th July 2005 First noticed SiSU   ⁴⁹³  in Debian Sid.

24^th July 2005 Further consideration of what is necessary to generate docbook SGML/XML valid output, initial concern is with mapping document structure. Will look into, but no time-scale, priority medium:

docbook valid output version? (no promises, distant future perhaps) comments on first step that is necessary

major (first) step is to get structural mapping from sisu to docbook, sisu to docubook structure mapping is not straightforward. Variations are fairly complex as the structure provided by sisu is different from that of docbook, in sisu there may be regarded as two sets of structural headings:

A. segment headings (book, section, part ... etc.) 1 to 3

B. text headings (article, section etc.) there are 3 levels (1 to 3 but) numbered continuing the sequence from 4 to 6

A. Segment headings variations to consider:

i. not all top levels are necessarily used, (where unused should be compressed, ie skipped and other levels moved up) and

ii. sometimes are used in a way that is not necessarily connected with structure e.g. (as in for author/creator, where by Dr X is used for level 2 or 3) - in this event ideally, no docbook level will be made

iii. and sometimes level 1 is used for parts of a document rather than document title e.g. cisg - in this event docbook should use title, (or create a wrapping level to represent the complete document)

B. Text headings follow a more strict structure (not all are necessarily used, but this matters less as are not followed by other levels). Document if with numbered headings is usually numbered from level B1 aka level 4 down (though sometimes parts belonging to higher up levels also have numbers, and it is possible to start numbering at a preceding level)

Possible solution

(a) start with title, not level 1, this wraps the document in its title, and permits use of level 1 for parts

(b) a header with an explicit mapping for docbook, explaining what should be mapped to what (consider later), and

(c) an option to place a marker within heading that results in its being omitted from having its own docbook level

(d) where level not used squeeze levels, omit for docbook and continue structure without

conclude

will result in some, though minor metaverse change to recognise

(a) headings in levels 1 to 3 that are not part of structure, (to be used for example where the author's name is used for level 2 or 3, possibly also where level 1 is a repetition of the document title, and level 2 the subtitle provided?) and

(b) possibly a new optional header for docbook mapping, which may follow later.

August

8^th August 2005 sisu 0.18.6

Will start placing document digests online, to get an idea of what this looks like see SiSU digest.txt   ⁴⁹⁴  may rename md5.txt,   ⁴⁹⁵  There is some polishing to be done the information provides several possibilities.   ⁴⁹⁶ 

Ratchet removed, as is unused and un-maintained.

9^th August 2005 0.18.7, removed some old conflicting license provisions, and fixes an md5 digest tree bug.

12^th August 2005 Complying with a Debian Developer's request that I get a GPG key:

  ID =
    BD76E77F

  Key fingerprint =
    F899 5A87 C648 3F38 5107  79F1 B97B 7C4D BD76 E77F

  -----BEGIN PGP PUBLIC KEY BLOCK-----
  Version: GnuPG v1.4.1 (GNU/Linux)

  mQGiBEL8fs4RBADNNhIsG6i13rDL4JLTgqy0fzcx8oJ9E/xDKeQo/5pqs18FvyKo
  VsHdkUrPptZgxowalPi2aWvZgdi56yUFt295pNedMNXfTnwkVF/9PSA3r4OTBFWw
  7NJYRUu2ZaSqtj3LpnBhsPpDqcl2g5Mry88mMLUeWB9OtsZE/QFVnoJqawCg4QSY
  IOtgdxZVJfibHbFA0/2iG3cD/A98l/BuRWF+kwNacj1q+D0glC/u9mvg56fnAzTL
  Osbhza/ebqvPgtFpP4PlIuvTcRZ9QsSJgmCTxNRPSKWPGqDmj7B7MDcPKUyJAPKB
  fSLm5v5BuKTCLl9jyihbOqIdTgF1hoBHxTyfYWpVVuhdFP1kahityDNGpRhSfTif
  doDLBACkB090RwyYBR589CMjF4miOo5m29ooQZO3tkR+6w5ny81H+qIHVx3TQZa9
  d2166YcxMOFLh2mXBcVNsF8gQpv8Otk+Oyu5FFKqJiGV6JkEPf/W+iXMdpvTrxtc
  rmU8cKagKxxBjue8PlI6bGusEjJXgmsTwOZB5WxCW4QQs49VoLQhUmFscGggQW1p
  c3NhaCA8cmFscGhAYW1pc3NhaC5jb20+iGQEExECACQFAkL8fs4CGwMFCQPCZwAG
  CwkIBwMCAxUCAwMWAgECHgECF4AACgkQuXt8Tb1253+R+gCg3KfzSGEmdWf01ryr
  QLm35LutXS8AoNVfXb3OB1BMph6bOSNJpIyZAOgstCdSYWxwaCBBbWlzc2FoIDxy
  YWxwaC5hbWlzc2FoQGdtYWlsLmNvbT6IZAQTEQIAJAUCQvyEaAIbAwUJA8JnAAYL
  CQgHAwIDFQIDAxYCAQIeAQIXgAAKCRC5e3xNvXbnf12SAJ44Tuntdnqv7lAGgRes
  KMFbHSt3YgCgjXkkHqY167kvDWmxrdAfupv7Ad65Ag0EQvx+2RAIAL23ShZnJJi/
  1un84eoUoQ/AaXmGAhYjVlYiJzMjTXBygS12TnpNkqibnnIqbn/6v1WFIgZNRhzS
  GzKo//MYtORdhB25UYe0YeOhZhyABLNSFKCye1ACBfDpJAFzc+kh7GjA1k2hnDsn
  zox49eJyubO1FxiXXYrKaSMOGkuhp/JaLEVHZZyY++Zq0TuS8t3bOmNkUbL5X5oZ
  n600yZzNfdQ8/5LjwIUvY9BDR+tHyypGyegDDmS5BVEjmW2d9ujDbVYRY6twA0q6
  6WSx2WqT/HzdvSirEiPp3ENwQAaTnGT0G5q2zmYtutb9Yd/V/5aQBMEkAXF/JT4H
  uw5ad3C6xNcAAwUH/jmC7wmvfA6Dqh9TfqDux9LzWbIny0NsgYTsE2Bkhg8kuHgb
  PeqpOeJRFYxQkhknzW1PZZVwtxmfFgWgvukW4fqGSPWSpWDaYL91zkx2HRMBG1Nb
  YwqNOLQnolRE9z/42b/M4oyIWhd0WTi6tZt5pMvpW5Dn7tOsnU3PlNM3AAv3Bjp9
  QBki1KRjZC7PdKnQa9KSwvr4hgvvBlT4Is9131vmAZhu41Ni8UORmoq5e/lqd5O/
  0QuySo+ZJEyLQuulfK19NPALZM+fu6/3dUHAu4D27MafYdQNqY3Xks4mquQ2p95k
  OhNer8Yu7SSyBCIO/8V71UhkgyUdG7fMl0BzqPmITwQYEQIADwUCQvx+2QIbDAUJ
  A8JnAAAKCRC5e3xNvXbnfzpwAJ0ZKPoWuWrx8K/zGSIRRj7ZRbf6KgCgji13BXoF
  tJaIUHauen7hNIoUQXo=
  =6US6
  -----END PGP PUBLIC KEY BLOCK-----

16^th August 2005 "Sphinx or Robot" by Leena Krohn   ⁴⁹⁷  added to the document examples   ⁴⁹⁸ 

Sphinx or Robot by Leena Krohn

September

6^th September 2005

sisu-0.23.0 iñtërnâtiônàlizætiøn a slightly more extended character set, using utf-8,  ⁵⁰⁸ 

! # $ % & \ () * + , - . / ' 1 2 3 4 5 6 7 8 9 : ; < = > ? @

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ]

< - ~ t ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ² ³ ' µ ¶ * ¸ ¹ º » ¼ ½ ¾ ¿

À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ÷ ø ù ú û ü ý þ ÿ

NOTE, sisu markup should files be saved with Unix line endings, and in utf-8 or iso8859.

7^th September 2005 sisu-0.23.1 generator of naive example search form for sqlite and pgsql databases added.

7^th September 2005

sisu-0.24.0   ⁵⁰⁹  released incorporating changes from development branch 0.23  ⁵¹⁰ 

21^st - 30^th September 2005 Ruby 1.8.3 released 2005-09-21 breaks SiSU, segfaults. Fixed in ruby 1.8.3 (2005-09-29)

Temporary solution was an emergency build of SiSU sisu-0.27.0 which did not use syck/yaml, (and that as a result was without optional user configuration and sisu version information). Thanks why_ and mfp for feedback.

Ruby segfault *fix* in ruby 1.8.3 (2005-09-29). Thank you TS/Guy Decoux.

New Ruby build not yet uploaded to Debian unstable. I rather hope it happens soon,...

Awaiting business as usual with Debian::Ruby::SiSU.

October

7^th-8^th October First look at Open Document format, downloaded Open Office 2 from Debian experimental. See October 18^th for examples. Omitted object character numbers from output, consider option to incorporate them, possibly in equivalently numbered notes for each object. See October 18^th.

12^th October Moved laptop from 30G to 80G hardrive.

Web enjoys year of biggest growth   ⁵¹¹ 

13^th October Ruby segfault on sisu  ⁵¹²  reappeared in ruby 1.8.3 (2005-10-12) and was again removed by TS:Guy in ruby 1.8.3 (2005-10-13).

13^th October Information Week Report: Linux is Cheaper, More Reliable and Secure, Encourages Innovation - October 13 2005   ⁵¹³ 

14^th October Started using Mercurial,   ⁵¹⁴  in addition to Darcs   ⁵¹⁵  Love them both. Mercurial seems to be more generic in its potential use space. Looks as though I will be doing micro (and probably macro) version control management in Mercurial, and package build releases with Darcs. Darcs (debian) build package is extremely convenient. Mercurial control will include Darcs repositories. Am extremely satisfied with both.

18^th October First pass at opendocument format included in sisu-0.26.4, sisu dumps some form of open document, though support is partial,  ⁵¹⁶  based on a couple of days look October 7-8^th Generated open document samples for book sample page.

21^st October Debian ruby-1.8.3-2 stamped with original release date (2005-09-21) fixes ruby-1.8.3-1 breaks of SiSU.

30^th October Test run on a few additional Project Gutenberg texts, transforming Gulliver's Travels to SiSU markup took 10 minutes:

"Don Quixote", Miguel de Cervantes [Saavedra] - translated by John Ormsby, PG Etext 996

"Gulliver's Travels", Jonathan Swift - transcribed from the 1892 George Bell and Sons edition by David Price - PG Etext 829

"Alice's Adventures in Wonderland", Lewis Carroll - PG Etext 11

"Through The Looking-Glass", Lewis Carroll - PG Etext 12

"Alice's Adventures in Wonderland" and "Through The Looking-Glass", Lewis Carroll - PG Etexts 11 and 12 composite text

November

November

ODF Open Document Format   ⁵⁸⁰ 

The articles that follow are themselves mostly links to other articles and comments generated by them, which i expect to be of less interest, but you never know.

Massachusetts' CIO Defends Move to OpenDocument   ⁵⁸¹ 

When Open Standards Really Matter - The Katrina Factor   ⁵⁸² 

MA Chooses OASIS OpenDoc XML as Office Standard - Requesting Comments   ⁵⁸³ 

It's Final - MA Goes With Open Document   ⁵⁸⁴ 

Comments on the Massachusetts Decision   ⁵⁸⁵ 

More OpenDocument News   ⁵⁸⁶ 

Tim Bray ["inventor" of XML] on Implications of OpenDocument Format   ⁵⁸⁷ 

Massachusetts Finalizes OpenDocument Standard Plan   ⁵⁸⁸ 

Massachusetts Explains Legal Concerns for Open Documents   ⁵⁸⁹ 

OpenOffice.org 2.0 Released   ⁵⁹⁰ 

OpenOffice.org 2.0 Released   ⁵⁹¹ 

MS Responds to MA/OpenDocument Announcement   ⁵⁹² 

More on Massachusetts - MS Talking Points & an Answer from Sun's McNealy   ⁵⁹³ 

Open Source In Public Sector Meeting Opposition   ⁵⁹⁴ 

Microsoft Says Maybe to ODF   ⁵⁹⁵ 

Search Google for "OpenDocument" - news stories on Open Document Format   ⁵⁹⁶ 

Answering Microsoft: Comments on Microsoft's Letter to MA ~ by David A. Wheeler, Comments on Microsoft's Letter to Massachusetts   ⁵⁹⁷ 

3^rd November Flirted with RCS as a lightweight alternative to CVS for tagging document versions. Both use the same Id tag, ... Back to CVS for this purpose (and yes Mercurial and Darcs for development)

added in debian:control:suggests, RCS as an alternative to CVS.

4^th November A character encoding problem of a user of SiSU led to my using lv. SiSU has become more demanding in requiring utf-8 for markup documents, this makes development a lot easier. The "problem" was converting from ISO-8859 English text, the solution

added to debian control, suggests lv

7^th November Added a manifest  ⁵⁹⁸  page page for document output and metadata. The like was previously incorporated into various presentations of the document including the html document information / metadata page, and an equivalent LaTeX / pdf output. However it makes sense to have a control page, that refers to other available output and presents the documents metadata.

8^th November Trying RCS again. For document version numbering CVS is more straightforward, but would like to have an idea of how RCS works with Mercurial. The current CVS snapshot for (sisu documents and a few test documents) is at:

cvs Id: sisu_chronology.s3,v 1.140

cvs Id: sisu_description.s3,v 1.43

cvs Id: sisu_download.s3,v 1.313

cvs Id: sisu_examples.s3,v 1.61

cvs Id: sisu_license.s3,v 1.22

cvs Id: sisu_provisional_patent_application_200408.s3,v 1.21

cvs Id: sisu_release_announced.ralph_amissah.20050104.s3,v 1.9

cvs Id: sisu_standard.s3,v 1.22

cvs Id: sisu_technical_info.s3,v 1.41

cvs Id: sisu_todo.s3,v 1.22

cvs Id: alices_adventures_in_wonderland.lewis_carroll.s3,v 1.5

cvs Id: autonomy_markup0.s3,v 1.24

cvs Id: autonomy_markup1.s3,v 1.16

cvs Id: autonomy_markup2.s3,v 1.16

cvs Id: don_quixote.miguel_de_cervantes.s3,v 1.2

cvs Id: eu_contract_principles_parts_1_to_3_2002.s3,v 1.10

cvs Id: free_as_in_freedom.richard_stallman_crusade_for_free_software.sam_williams.s3,v 1.30

cvs Id: free_for_all.peter_wayner.s3,v 1.15

cvs Id: gpl2.fsf.s3,v 1.14

cvs Id: gullivers_travels.jonathan_swift.s3,v 1.2

cvs Id: through_the_looking_glass.lewis_carroll.s3,v 1.5

cvs Id: un_conventions_membership_status.s3,v 1.10

cvs Id: war_and_peace.leo_tolstoy.s3,v 1.2

27^th November Finally introduced to pbuilder

29^th November Debian packaging for SiSU changed, broken into smaller packages, for better control over dependencies

       sisu
       sisu-composite
       sisu-examples
       sisu-pdf
       sisu-postgresql
       sisu-sqlite
       sisu-termsheet
       sisu-texinfo
       sisu-vim
       sisu-webrick

man pages are included

December

10^th December Check SiSU default settings against a couple of Debian documents:

28^th December Multilingual version of document, model conceptually sorted, and an implementation in sisu-0.35.0 released early today, sample output and markup:  ⁶¹⁷ 

SiSU document sisudoc.tgz becomes a sisupod, with multiple language markup versions zipped.

years end Technology trends of interest in 2005

Business Week claims this 2005 was a [yet another] breakthough year for Open Source   ⁶²⁰ 

Strong growth for Debian   ⁶²¹  interestingly the statistics omit last years most successful newcomer, Ubuntu, which is Deban based, and would no doubt impact a bit.

Ruby book outsell Python at O'Reilly   ⁶²²  according to O'Reilly, interestingly they publish a number of Python books, and have one now out of date Ruby book of their own. (the significant jump fuelled by interest in Rails for web applications).

an IBM DeveloperWorks article on Ruby comparing it favourably with Java   ⁶²³