2.1 Preparation of the directories – The files containing the documents must be in a unique directory, with two different subdirectories named divers and resume. The main directory will contain the compacted Postscript files containing the ArXiv articles. The directory divers will contain the other documents. The directory resume will contain the summaries of ArXiv articles.

Another directory, containing some files produced and used by STR must also be created (it may be a subdirectory of the previous main directory). It must contain a subdirectory named Auteurs. It must also contain the main page liste4.html of figure 1 (which can be easily modified to use another language).

Finally, another directory must be created to contain the temporary files used when viewing items.

 2.2 The initialization file – At startup, STR and the visualization program read the file .listerc3 which should be in the main directory of the user. It contains 13 lines giving the following information:

1. Line 1: name of the directory containing the files produced by STR and some files used by it – Example: ~/Articles/DB
2. Line 2: name of the directory containing the documents – Example: ~/Articles
3. Line 3: – name of the suffix of the compacted articles from ArXiv Example: ps.bz2
4. Line 4: name of the suffixe of the articles in Postscript from ArXiv, once they have been decompacted – Example: ps
5. Line 5: name of the suffix of the articles written in dvi and compacted – Example: dvi.bz2
6. Line 6: name of the suffix of the files of articles in written dvi, once they are unpacked – Example: dvi
7. Line 7: name of the temporary directory, followed by a prefix – Example: ~/tmp/XXX . Here “~/tmp” is the name of the directory that must be created (cf. 2.1).
8. Line 8: name of the decompression program – Example: bunzip2
9. Line 9: name of the program used to for view ArXiv articles in PostScript – Example: okular
10. Line 10: name of the program used to view the other articles in PostScript – Example: okular
11. Line 11: name of the program used to view the other articles in dvi – Example: okular
12. Line 12: name of the program used to view the other articles in djvu – Example: djview
13. Line 13: name of the program used to view the other articles in pdf – Example: okular

In the previous examples, the file .listerc3 will contain no the lines

~/Articles/DB

~/Articles

ps.bz2

ps

dvi.bz2

dvi

~/tmp/XXX

bunzip2

okular

okular

okular

djview

okular

 2.3 Formats of the documents – The Arxiv articles – They must be in Postscript and compacted, their name must be of the form ZZZZZZ.suffix, where ZZZZZZ is a name given by the user (for example NameAuthor-05), and suffix is given in line 3 of the initialisation file .listerc3 (cf. 2.2, hence in the preceding example, it should be NameAuthor-05.ps.bz2). The file containing the resume of the article and saved from Arxiv should have a name with the same prefix ZZZZZZ, followed with .html (hence in the preceding example it will be NameAuthor-05.html).

 2.4 Formats of the documents – The other texts – The other texts (non ArXiv) can be in compacted Postscript, in compacted dvi, in pdf or in djvu. The file names can be anything, except the compacted Postscript and dvi files, which must conform to the indications contained in the initialization file.

 2.5 Compression and decompression – The same program must be used to compact the files in Postscript or in dvi. It should produce files with suffixes corresponding to what is indicated in the file .listerc3 (cf. 2.2).

 2.6 Liste of the other articles – A file containing the of list non-ArXiv articles and information about them must be created and updated each time that a new article is added. The name of this file is liste_div.lst, and it must be located in directory of files produced and used by STR, given in the first line of the initialization file .listerc3 (see 2.2). Each article is described by 4 lines of information, and each of these blocks of 4 rows must be separated from the next by a blank line. The 4 rows contain the following information:

1. Line 1: name of the file containing the article
2. Line 2: names of the authors. Voir plus loin.
3. Line 3: title
4. Line 4: format of the article. It is ps, dvi, pdf or djvu.

If the name of an author has multiple parts, they must be separated by the character # instead of spaces, otherwise these parts will be considered as different authors names.

Example:

 2.7 Alias – Some names of authors (those of articles from ArXiv automatically recovered from abstracts) can appear written in different ways, especially if they have accents, and be counted as names of distinct authors. To avoid this inconvenience, it is possible to indicate in a file the different names of the same author. This is the file alias.lst, located in the directory of files produced and used by STR, and mentioned in the initialization file .listerc3 (see 2.2). This file is made up of lists (one for each author) separated by a line containing the single character # . The first row of a list is the real name of author, others have other names that can appear.  Note: the  main (first) author name of each list must be used by some article.

Exemple :

Dupré

Dupre

Dupr

#

Möppe

Moppe

Moeppe

Mppe

 2.8 Practical implementation – With the parameters of the initialization file .listerc3 given in 2.2, one must proceed as follows to add a new article:

To add a new article from ArXiv, download the article in PostScript (and compact it) as well as his summary, their name must have the same prefix, for example Article23.ps.bz2 will be the name of the compacted Postscript file and Article23.html that of its summary. The article will be placed in the directory  ~/Articles  and the summary in  ~/Articles/resume.

To add another article from another source, it will have to be in one of the supported formats (compacted Postscript, compacted dvi, pdf or djVu). It will be located in the directory  ~/Articles/divers. The file  ~/Articles/DB/liste_div.lst  must be then completed with four lines describing the article, as shown in 2.6.

To update the liste of articles, one must then run the program STR, and give it the command all, followed by exit:

 2.9 The command ’clean’ – It can be given to STR to delete all the files produced by it (but not the documents !).