automated profile generation

Reinhard bookreader at
Wed Mar 24 09:08:52 PST 2004

Thank you very much, James, for your attention and that piece of info!
I tried to get rid of the working process by reading the mails from the 
archive, but sure - I did not read all 4800 mails - so may be I missed so 
important steps.

On Wednesday 24 March 2004 16:12, James Robertson wrote:
> Thanks for the info and comments Reinhard.  I too would like to see
> search_replace simplified and made more powerful.  
> I saw that Kevin has that in his plan and that is good enough for me.

Don't take me wrong. I didn't want to be importune, I just wanted to correct 
some missunderstandings - I sometimes feel missunderstood 
... may be cause I'm not english by nature.

> I also know that package dependencies are on the list as well to make 
> nALFS act kind of like a package manager.  Well built profiles with all the
> dependency info in them, the tool would be able to use the logs of already
> installed packages to determine the order to install the new stuff.
> This is in the roadmap.  

Thanks for that info.

> With your help on pulling the information out of the XML source directly, 
> I think, could help Thomas and Jamie.  I'll let them make that
> determination. 

Well for me the most valuable property is sparetime. I don't like wasting my 
time and I think this is true for all - editors, users, what ever.
So IMHO editing profiles, that could be generated is wasting time.
Having to read a whole book, where a tool could give you a path - is wasting 
I like to decide myself, when I just like to use something and when I like to 
learn more about something.
That was the reason for me, to write a tool that pulls the information out of 
the book and gives me a path for installation.

> I have no issues with perl scripts to add functionality to the tool.  

No problem at all. It was my choice to learn perl and the script was a good 
example. Monitoring the list I read about the problems with the new 
xml-format and as the buildin tools died with the cvs-versions of the book, I 
wrote my own xml-engine.
(I have some experience in C++, so if you like to integrate some features in 
nALFS - and like me to do some porting - let me know.)

Checking the errors I could find bugs in my script, but I could also find 
"suboptimal" content in the book.

I think, it should be possible and desirable, changing the book-content so 
that the HTML-version meets the perception of the editors and the XML-version 
contains all information neccesary for automated processing without any 
manual editing of profiles or whatever beside the book.

Cheers Reinhard

More information about the alfs-discuss mailing list