No previous Up Next

Site maintenance

On this page, I give some information on how I maintain this site.


For editing, I use a variety of text based editors, such as:

Internal link checking and uploading

From my HTML-to-LaTeX program, I created a HTML syntax and link checker, which also determines which are the files that have been modified and need to be uploaded. Furthermore, it also generates the site statistics as found on my main page, and the file with all the broken links. The program is called from the chkhtml.bat batch-file. This batch-file is executed from the directory that contains all the files that make up this web site. It does the following things: For uploading, I start-up the WS-FTP program in the ftp sub-directory. After, all files are uploaded, I press the dirinfo button for the remote directory, and save the file as compare\ls.txt. Of course, I also empty the ftp directory.

External link checking

I checked all my links to external pages with the free Xenu link checker. Instead of letting this program scan all my files, I usually let it execute on exrefs.txt, which is generated during the internal link checking. The program can generate a report, and the result can be saved to a tab separated file. The results I save in the file compare\links.txt.

The report I process with the program exredirect.c, which produces a file with all the redirections as mentioned in the report. Each line contains one redirection separated by a tab-character. I edit this file by hand to see if any of the redirections really represent an error (Some site redirect not-found errors to a special page). A file can be marked as an error by replacing the string after the tab by something starting with the # character, for example, with "#not found". I save this file as compare\redirects.txt.

After, I have done this, I run the program chkextrefs.c which reads the contents of the files compare\dir.txt, compare\links.txt and compare\redirects.txt. The program will scan through all the HTML files, and make the neccesary changes. Any broken link is changed into a reference to Broken.html?some error|original link. Whenever a HTML file need to be changed, the changed file is placed in the converted sub-directory. The original file is thus not affected. After a sanity check, the files from the converted sub-directory can be placed back in original directory.

File exchange

For the exchange of files between the various computers on which I do my editing (currently only ester and the Toshiba laptop), I make use of a MS-DOS floppy. The reason behind using the MS-DOS filename format, is because the Toshiba laptop runs under MS-DOS.

I wrote a small C program to support the file exchanges. For the mapping between the long filename format and the "short" MS-DOS filename format, it makes use of a file with the name content.l2s which is stored on the exchange floppy. When the program is ran, it compares the files on the floppy with the files on the harddisk based on the mapping described in contents.l2s. Based on this, it produces a batch file with all copy statements for transfering the files. I Usually edit this batch file first, before I execute it, to remove the files that I do not want to transfer.

The program is called by the batch-file syncro.bat, which is executed in the home directory of all the files that make up this site. This batch-file makes a directory listing of this directory in the file compare\dir.txt, and a directory listing of the floppy in the file compare\adir.txt. The the syncro program is called and the output is directed to the batch-file syncflds.bat, which is then edited with notepad. On my desktop, I have a short cuts to the syncro.bat and syncflds.bat batch-files. In the past, I have used more advanced programs, and I also experimented with a VB program, but this simple set-up works the best as it gives me full control over the whole file copying process.


Although I do not focus on graphics on my web pages, there is enough of it. Also for my graphics, I use some simple free programs. These are: (Annabel used QV to make some drawings.)

Scanning text

The best way to scan some text for web display is: For an example, look at this image, which is only 21.812 bytes in size. If you zoom in on these kind of scans, you notice the amouth of grey pixels in the image, and that the text actually become less readable. If you have only a 300 dpi scanner, you can use the following recipe (thanks to Henk Vink):

Creating transparent GIF's

For creating transparent GIF, I use my own program program (8 Kbyte MS-DOS executable), which I wrote myself after seeing a small perl script doing the trick. If you run the program it displays the color table. The program allows you to give you an index of the color that should be made transparent and to specify an output file.