Vanitech news and report: Difference between revisions

Revision as of 13:26, 23 December 2009

Currently we have a few projects we are working on:

The improvement of the tools by which we compile quotes in Vaniquotes

The Vaniquotes compiling system works by taking portions of text from the Bhaktivedanta Vedabase and formats them as "vaniquotes" in order to be inserted into pages.

We have finished working on the 2.0 version that presents a rich text (WYSIWYG) interface with automation on: 99% link (to Vanisource) creation accuracy, cited verse in quote detection and formating, text formating according to its attribute (heading, translation, purport or conversation), spacing in text and between quotes, insertion of table of contents and placement of quote in page according to its (semantic) nature: BG, SB, CC, Other books, Conversation, Lecture, Letter and their subcontents (BG -> BG Chapter 1, Other Books -> Nectar of devotion, Lectures -> Bhagavad-gita As It Is Lectures, etc).

The 2.0 version has been released and tested thoroughly during the 2009 Srila Prabhupada's Vanimarathon but of course there is always some bugs to fix and room for improvement:

Statistics have been kept for all compilations created. We want to extend this by having also statistics and goals automatically populated by compiler. With the infrastructure and systems we have in place it shouldn't be too hard to implement this, since it was already in mind at the time of the Statistics system creation. It will populate a list of the compilers with their goals and the daily amount of quotes (and pages) compiled. Compilers will also be able to set their own goals and follow up on them.

Automation of submission of completed pages for Editors to be notified and improve the communication with compilers on compilation issues and improvement.

Creation of pages is vital and we wish to extend it to very serious and committed compilers who will have even more tools do this wonderful and very academic and research oriented service. The new feature will be auto-suggestion of synonym and antonym categories for new pages based on the pages categories and expressions list. This will help the researcher (compiler) find even more ways to study and populate themes and topics based on Srila Prabhupada's teachings.

The person currently in charge of this project is Acyutananda Caitanya Das.

The development of Vanimedia: The repository of video, audio and pictures of His Divine Grace Srila Prabhupada

A dynamic, very appealing and easy to use gallery where users/visitors can browse and access the huge repository of media.

video - 50 hours

audio - 1700 clips

pictures - 20,000 pics

We are still in search of qualified and enthusiastic devotees who would like to build the most fun and dynamic way to browse through Srila Prabhupada's pictures, audio and video.

The indexing of Srila Prabhupada's audio and video

A searchable media system that will be tied to Vaniquotes and Vanisource, in order to supply the audio (and possibly video) of vaniquotes from within a Vaniquotes page. We will also have audio inside Vanisource Conversations and Lectures so we can listen to the entire recording or portions of it.

As far as our knowledge and research goes, we envision to either:

find the software or a qualified person to write the software (with currently open source ASR - automatic speech recognition - libraries) to "tag/index" the audio (and possibly video) in the way that google labs is doing in their beta version of Gaudi (searchable audio); or
create a system to "index" audio, in a similar way we are compiling text in vaniquotes: by engaging hundreds of volunteers in a collaborative network.

Based on extensive research done by Bhakta Robert prabhu on speech recognition's immaturity and (in)accessibility for us humble servants, we are concluding that the semi-automated indexing system presents a more feasible alternative to automated indexing of Srila Prabhupada's media through ASR. We are also very enthusiastic about this model since it will present devotees with a new opportunity to study and research Srila Prabhupada's vani, now through hearing opposed to reading as done in Vaniquotes.

This indexing tool will be accessible through Vanisource and wherever the audio is available, and the reader is interested in rendering vaniseva in this way, she will have the opportunity to do it in a very simple way. Once the audio in a page (Lecture or Conversation) has been indexed, it will become immediately available in that Vanisource text and in EVERY quote in Vaniquotes that was extracted (through the process of compiling) from that lecture or conversation. Imagine how amazing and fun it will be to audibly browse through a Vaniquotes compilation, and listen to the audio of every lectures and conversations directly in a Vanisource page!

Searchable media was in hold for some time but now it will become a reality since we count with very qualified developers who can make this happen.

How is this related to the Vanimedia petal? Well, Vanimedia is our catalog of media so we want to pull the media from this main media repository (avoiding duplication of resources) and eventually we will implement this search system inside the Vanimedia experience for everyone to make "audio and video searches" in Srila Prabhupada's media.

The persons currently in charge of this project are Bhakta Robert Dinah, Bhakta David Roncancio and Acyutananda Caitanya Das.

Vanisource clean-up system and import of BBT earlier editions

We are committed to provide the most authoritative repository of spiritual knowledge based on the "Vani" given by His Divine Grace A.C Bhaktivedanta Swami Prabhupada. For this we have imported all of his writtings and teachings in the text form into the Vanisource petal extracted from the Vedabase software. Unfortunately we have inhereted many errors in the texts that need to be corrected and we believe that responsibility lies in all of us not just the publishers and developers of Vedabase (BBT). We have one great advantage compare to the system that may be offered by the publishers: We have already in place a bulletproof collaborative system (Krsna bless mediawiki), in which we can make this corrections and have instant benefit of it.

We will create a very easy system by which registered users can submit corrections (misspells, characters missing, punctuation, etc) of the text which will be immediately reflected (after revision) in the Vanisource text and the Vaniquotes that were created with that text. In this way we hope that the publishers will use the corrected version made by hundreds of volunteers whose sole interest is to propagate the knowledge they were given in the quality it deserves.

The vanitech team has also been busy importing the previous editions of the main books written by His Divine Grace. This is done with the purpose to have a easy and accessible way to read and compare different editions of the texts. We are solely interested in academic and devotional research and have NO interest in politics or conspiracies, reason why we endeavor to provide the facts.

There is already available the 1972 version of the Bhagavad-gita which can be viewed and compare with the current (1983) edition by visiting any BG text in Vanisource. We will extend that to other books as time permits. We hope this encourages devotees and ISKCON authorities to make their own judgment and learn about the corrections made in the latest editions.

Another improvement we wish to have in our Vanisource is the accessibility to the original Sanskrit (and Bengali) Devanagari in every text. Because our source is the Vedabase, and at the time the Vedabase was made (making Srila Prabhupada's books in digital form) the standards for Sanskrit Devanagari were not so established as they are today (This is of course merely a conclusion drawn from our vanitech team since haven't had the pleasure to hear it directly from the developers of the Vedabase software). The Vedabase created their own font with their own encoding for the sanskrit devanagari, making it only accessible to those who have the font installed in their computer. Meaning that if you were to copy and paste a portion of the text to a text (word) document or an email (either a client software or the more popular way the internet browser) your text will appear in a strange code unless you have the font installed.

For this purpose we have created a tool that will allow us to convert, in the same way we converted the "romanized sanskrit" in every Vanisource verse, to a standard universal (unicode) encoding that nearly all personal computers can read (display properly). The difference between the conversion of the romanized sanskrit (Balaram to Unicode) is that since that encoding doesn't meet any standard encoding known to us, we have to use a semi-automatic tool that will allow a volunteer with a good understanding of sanskrit, to verify and correct our version of the text. There are already a couple of volunteers, Archana Dasi and Bhaktin Matea who will start working on the devanagari conversion in the Bhagavad-gita and once the tool is refined through this beta process then we'll continue with Srimad-Bhagavatam and other books. You are welcome to join them in this nice effort.