1st 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.
3rd January 2005 Guess this should have been looked at earlier, added:
4th January 2005 SiSU release announced. 412
There have already been many subsequent updates - for the current release details see:
11th January 2005 A more complete bundle has been posted.
27th January 2005 SiSU separate Stable and Development Branches are needed and will be introduced shortly, based on a version numbering policy. 417
31st 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 Sabaki: version 0.1.2-19 of 2005w05/1 (20050131) & Ruby: ruby 1.8.2 (2004-12-25) [powerpc-darwin7.7.0]
SiSU Sabaki: version 0.1.2-19 of 2005w05/1 (20050131) & Ruby: ruby 1.8.2 (2005-01-10) [i386-linux]
3rd 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 that release is withdrawn as it does a bit too much in root.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.
4th February 2005 SiSU installation starting to work more smoothly, tested on Debian and Fedora using Rake. Thanks G.A.
5th 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 4th, 2005 :-) ]
12th 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
15th February 2005 SiSU Sabaki 0.1.4-0 (0.1.4 is stable branch)
16th February 2005 Confirmation today that basic SiSU install with Rake has no problems on Gentoo, thanks G.A. Current list Debian, Fedora, Gentoo.
21th 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.
2nd March 2005 Some work done in development branch on internationalisation.
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.
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.
Markup Syntax summary:
SiSU development dates back many years. (documents need rewriting, they predate the release of SiSU).
man pages, interactive help provided.
10th 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
14th 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).
27th 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.
10th 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).
12th 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.
13th April 2005 Re-introduce sqlite module.
16th April 2005 Debian Sid includes ruby1.9, first look, for SiSU, that's how lazy Debian keeps me, may Debian keep it that way.
19th 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
24th April 2005 New branch 0.6 ready.
26th 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.
28th April 2005 New branch 0.8
3rd 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
6th May 2005 An early Debian package of SiSU sisu_0.10.1-1_all.deb and sisu-examples_0.10.1-1_all.deb
8th May 2005 Early Debian package of SiSU, paths sorted: 443
which maps to:
17th May 2005 Headway with Debian packaging, much automation. Still detail to sort, and have not checked all dependencies.
20th 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:
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.
21th May 2005 Changed package directory directory structure, and references to it.
17th 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.
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 4th, 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).
19th 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.
21th 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.
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.
Debian package: ‹http://www.jus.uio.no/sisu/SiSU/download.html#debian›
[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
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.
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 - 2nd August 2005 Kimito, Finland, mostly offline
14th 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 20th July 2005, but need to be consistent about content of objects on which md5 digest is made, so relates directly to sql clean.
22nd 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.
24th 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)
(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
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.
8th 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.
9th August 2005 0.18.7, removed some old conflicting license provisions, and fixes an md5 digest tree bug.
Key fingerprint =
F899 5A87 C648 3F38 5107 79F1 B97B 7C4D BD76 E77F
-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.4.1 (GNU/Linux)
-----END PGP PUBLIC KEY BLOCK-----
6th 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.
7th September 2005 sisu-0.23.1 generator of naive example search form for sqlite and pgsql databases added.
7th September 2005
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.
7th-8th October First look at Open Document format, downloaded Open Office 2 from Debian experimental. See October 18th for examples. Omitted object character numbers from output, consider option to incorporate them, possibly in equivalently numbered notes for each object. See October 18th.
12th October Moved laptop from 30G to 80G hardrive.
13th 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).
14th 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.
18th 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 7-8th Generated open document samples for book sample page.
21st October Debian ruby-1.8.3-2 stamped with original release date (2005-09-21) fixes ruby-1.8.3-1 breaks of SiSU.
"Alice's Adventures in Wonderland" and "Through The Looking-Glass", Lewis Carroll - PG Etexts 11 and 12 composite text
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.
3rd 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.
4th 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
7th 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.
8th 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
27th November Finally introduced to pbuilder
29th November Debian packaging for SiSU changed, broken into smaller packages, for better control over dependencies
man pages are included
10th December Check SiSU default settings against a couple of Debian documents:
28th 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
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).
401. 2004 ended on a dastardly note.
412. sisu_0.1.0-9 ‹http://www.jus.uio.no/sisu/SiSU/changelog.html#first_release›
413. Reproduced with the kind permission of author and artist Leena Krohn, ‹http://www.kaapeli.fi/krohn› Her online works include Tainaron ‹http://www.jus.uio.no/sisu/tainaron.leena_krohn.1998› and Sphinx or Robot ‹http://www.jus.uio.no/sisu/sphinx_or_robot.leena_krohn.1996› which can be found in the book sample section ‹http://www.jus.uio.no/sisu/SiSU/examples.html#sample›
417. 26-27 January 2005, introduced a feature and with it a bug that persisted through a couple of releases. Bugs fixed 0.1.2-17 feature intact. (for further information, see release notes in download section.)
418. Thanks for looking O-E P
instructions are provided in development section:
428. a relatively small extended character set, currently available (LaTeX/pdf,html,XML,Postgresql): ¿ 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 år æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ø ù ú û ü þ ÿ
429. ¢ £ ¥ § © ª « ® ° ± ² ³ µ ¶ ¹ º » ¼ ½ ¾ × ÷
430. European A4, A5, B5 and U.S. letter, legal
431. removed colour for visited urls
432. default red/ruby coloured composite, and blue/saphire (the latter found at darcs wiki)
437. An interesting possibility to consider, is to permit sisu to get the composite parts of a document from a remote url source via http, using thlnk syntax which permits this.
438. Reference, Old Syntax Examples
Old Syntax, example where endnotes follow paragraph, html for text highlighting
Old Syntax, example where embedded Endnotes are contained within paragraph, html for text highlighting
Old Syntax, Old Style Headers, and Old Style Endnotes follow paragraph, html for text highlighting
Old Syntax, Or again for the earlier paper
439. all SiSU objects being numbered sequentially
443. tested on i386 it should work on all Debian supported architectures.
sisu_0.10.5-1.dsc: cf5a87ca297a2b026ea5c06ac39be91e 3079976 sisu_0.10.5-1.tar.gz
Filed August 17, 2004, Application Number 60/602,290, Filing Recept OC00000001 37841 68
451. Rudimentary implementations have existed for as long, and have been used in the building of the web site lexmercatoria aka. Ananse, the International Trade Law Project and the International Trade Law Monitor; and major parts of the implementation have been complete since 2002, ‹http://www.jus.uio.no/sisu/diagram/sisu.chart.pdf›
452. ending up with a work environment based on debian/gnu/linux or unix, ruby, postgresql, vim, zsh, screen, ion etc. (plus broadband and a laptop with a high resolution monitor)
453. aka. Ananse, the International Trade Law Project and the International Trade Law Monitor.
454. and I had the benefit of a friend as attorney who worked pro bono
455. Filing a provisional patent application gave a year in which to decide whether to proceed with the patent process.
457. They had existing document management systems which he thought could benefit from the logic in the code demonstrated. In particular, they were interested in the document search possibilities, as their product was only able to return information on the documents which matched the search criterion.
SiSU when used with a database provides this at a "paragraph" (text or other object) level, and will return the documents matched together with the locations within each document of the match, or the documents matched together with the objects in which the match was found.
461. 2005w24/0 i.e. 20050619
464. corrected 20051112 from
Serialized information, Structured Units
470. this was implemented shortly afterwards, during the holiday that followed, making digests available for any further desired processing by other SiSU modules, generating them as output, and placing them into sql databases.
who provided this link ‹http://developer.skolelinux.no/dokumentasjon/peters_docbook_cookbook/›
477. no recent home page found
483. sisu_0.16.25: sisu_0.16.25.orig.tar.gz, sisu_0.16.25-1_all.deb sisu-examples_0.16.25-1_all.deb sisu-vim_0.16.25-1_all.deb
495. as the document may be published in different locations it is difficult to place link within the digests page back to the corresponding html index page
fixed bug in 0.18.7, had noticed that there was an important bug todo with garbage cleaning that appears in batch processing mode.)
496. e.g. it could be used to provide diffing information (though in its form now only the latest md5 tree is stored, may consider option to retain several, possibly by date); it could hypothetically be placed with a third party evidence of contents of a document at a given date without disclosure of the actual contents, and; being available for incorporation in relational databases can be used for search of identical objects.
508. tested so far on LaTeX / pdf output, line-breaking is not perfect
BBC, 10 October 2005
512. which first appeared in ruby 1.8.3 (2005-09-21) and was removed by TS:Guy in ruby 1.8.3 (2005-09-29)
Related to the fact that software Patents (in their current form) stifle rather than promote innovation (they are more important to prevent others from developing ideas than encouraging new ones, and sometimes blocking off wide areas of development/innovation); and are only meaningful to big business who can afford them at USD 10 - 15K a pop, and who can afford to defend them at multiples of USD 100K a time. In addition to all that it is usually akin to the patenting of logic (or maths)... all another big issue and can of worms for society.
516. e.g. images and tables are not supported for detail see changelog
The OpenDocument technical specification (pdf 700 pages):
581. ‹http://politics.slashdot.org/politics/05/11/01/1947225.shtml?tid=109&tid=219› Slashdot, 1 November 2005
582. ‹http://www.groklaw.net/articlebasic.php?story=2005091305273070› Groklaw, 13 September 2005
583. ‹http://www.groklaw.net/article.php?story=20050831202118904› Groklaw, 31 August 2005
584. ‹http://www.groklaw.net/article.php?story=20050923142231938› Groklaw, 23 September 2005
585. ‹http://www.groklaw.net/article.php?story=20050925165302314› Groklaw, 23 September 2005
586. ‹http://www.groklaw.net/article.php?story=20051006155004596› Groklaw, 6 October 2005
587. ‹http://slashdot.org/article.pl?sid=05/09/30/0632243&tid=185&tid=187› Slashdot, 30 September 2005
588. ‹http://slashdot.org/article.pl?sid=05/09/24/1317234&tid=109&tid=103› Slashdot, 24 September 2005
589. ‹http://yro.slashdot.org/article.pl?sid=05/09/06/0345224&tid=155&tid=185› Slashdot, 6 September 2005
590. ‹http://www.groklaw.net/article.php?story=20051020092331261› Groklaw, 20 October 2005
591. ‹http://linux.slashdot.org/article.pl?sid=05/10/20/1232218&tid=185&tid=106› Slashdot, 20 October 2005
592. ‹http://www.groklaw.net/article.php?story=20050906080305991› Groklaw, 6 September 2005
593. ‹http://www.groklaw.net/article.php?story=20050911201041458› Groklaw, 11 September 2005
594. Slashdot, 29 September 2005
595. ‹http://www.groklaw.net/article.php?story=20051016105739574› Groklaw, 16 October 2005
597. ‹http://www.groklaw.net/article.php?story=20051029212458555› Groklaw, 29 October 2005
598. originally named summary and summary.html was renamed SiSU manifest and sisu_manifest.html 2005-12-16 to reduce the chance of filename collision, or rather, to make a less likely to be wanted for regular use name a SiSU reserved filename.
617. Implementation relies on a ssiu file naming convention. Use of country codes for language will be changed or extended.
623. ‹http://www-128.ibm.com/developerworks/library/j-ruby/?ca=dgr-lnxw01RubyOffRails› IBM DeveloperWorks, 20 December 2005
‹http://developers.slashdot.org/article.pl?sid=05/12/24/202200› Slashdot, 24 December 2005
Eric von Hippel
Erik S. Raymond