Organizational Arrangement of the Project
Hana Velickova
Parliamentary Library,
Czech Parliament
1. Technical specification
1.1. Introduction
The detailed technical specification which accurately describes the way of document processing is a constituent part of each contract. In the Electronic Library you can find two basic types of parliamentary materials, debates and parliamentary documents. In spite of the fact that these parliametary materials are quite different their processing have a lot in common:
- All the texts are saved in HTML, in a "clean" code. It means that the HTML does not comprise Font Face, metatags etc. and information that is added to the HTML source file by various editors. Such HTML files are small and well legible by any web browser.
- Basic graphical lay-out of all the texts is also the same:
- normal, bold, italics and underlined types are kept
- spaceout types are converted into bold (due to the full text searching)
- titles are marked by bigger types
- centered texts are kept
- paragraphs are separated from one another by one iserted row
1.2. Debates
All the debates (debate is a verbatim report from a parliamentary meeting) from the whole processed period have nearly the same structure, e.g. the summary in the beginning, the program of the meeting, sometimes supplements of the debate and then concrete debates. The texts are saved into files on the basis of their structure, e.g. the summary, program and supplements are saved in separete files. Debates are separated into files of size 30 kB (each file is a logical unit, the text in the file ends with the talk of one speaker or with the end of a paragraph).
1.3. Parliamentary documents
Parliamentary documents are all the materials that are discussed on a parliamentary meeting. In contrast to the debates they differs one from the other very much. There are a lot of types of parliamentary documents, e.g. government or parliamentary bills, questions and answers, resolutions of commeettes, reports, budgets etc. that each have different graphical lay-out. Sometimes there is different lay-out of one type of parliamentary document even during one parliamentary session. Furthermore these texts include a lot of graphical elements that are not contained in debates:
- graphs, tables, copies of articles etc. are inserted into a text as the pictures in the GIF format
- if the signature is not legible or there is not known the author of the signature the signatures are again inserted into the text as a GIF picture
- logos, national emblem and other pictures in the headings are left out
- footnotes are inserted into the text in italics and brackets
Long parliamentary documents are again separated into parts of 30 kB.
2. Organisation of the cooperation with the firms
In spite of the fact that the detailed specification is a part of the contract sometimes only when processing the firms come across problems that were not solved in the specification (while preparing the specification we cannot go through all the pages and it is not possible to encounter all the problems). These situations we solve during processing and inform both the firms about results and new solutions.
The processed texts we get on CD-ROM. After the taking-over we start to check the quality which process is detaily described in the contract: two people read and examine accidentaly chosen parts of processed texts for the time of one hour. During this time they should read 60 pages (e.g. 60 characters oin 30 rows), e.d. approximately 100 000 characters. If they found more than 100 wrongly converetd characters (e.g. the error rate would be bigger than 0.1 % - 1.7 mistakes in one page) the order would be returned back for reprocessing. The guarantee period stated un the contract is one year.
The results are as follows: the average error rate of the first firm is 0.0015 % (0.27 mistakes in one page) and the second 0.008 % (0.15 mistakes in one page) that is in both firms more then ten times less than the limit. As you can see the quality of processing is very high and we are satisfied with both the firms.