First version of NAIS is out.

[Jens Rühmkorf]

Hello Thomas, hello everybody! 

Please excuse that the answer to your email took so long, but some of us
were on vacation.

On Thu, 15 Jun 2000, Thomas Gebhardt wrote: 

> maybe I missed something, but what does the acronym "NAIS" stand for?

What does "NAIS" mean?

a. "nais" derives from the french verb "naitre"="to be born", used as an
   imperativ "nais!" means "be born!". (The italian verb "fai" on the
   other hand has the Meaning "you make".) 

b. Additionally "NAIS" can be understood as an acronym for "network
   automated installation system". This was rather a matter of coin-
   cidence not choice, though ;) 

> To be honest, I was rather confused about your announcement about
> that code fork (especially since - as far as I can see - the authors
> of FAI and NAIS are stable mates). 

Well, this is nothing that happened "out of the blue". But indeed, it is
true that generally speaking such a code fork does not make sense.

> Having read the documentation of NAIS I got the impression that the 
> improvements and new options implemented by NAIS could be integrated
> into FAI, at least as optional features. Do you consider FAI to be 
> fundamentally flawed such that it is necessary to start a new project? 

When answering this question you have to distinguish the person(s) behind
FAI from the project itself. Of course we do not consider FAI to be
fundamentally flawed, otherwise the software and concepts Jens, Thomas
(Lange) and I developed were to be regarded as "bad" in the first place. 

> Well, both FAI and NAIS are GPLed, so there is no problem in exchanging
> code and concepts; but I still cannot understand why one cannot join the
> effort and make the ultimate installing system jointly. 

This is true, and NAIS is GPLed for good reasons. We want to give others
the opportunity to contribute to NAIS as authors to help build "the
ultimate installing system". 

To get back to this question:

> Do you consider FAI to be fundamentally flawed [...]?

We do think FAIL has its weaknesses. What we considered to be most
important we implemented for NAIS first:

- If you have been following the threads during the last weeks you will
  have noticed that FAI has problems installing some packages, sometimes
  the package will fail to install at all or it takes more time than it
  should (sendmail). The concept NAIS uses is much more robust, because
  packages are installed in a "normal" debian environment where all
  install scripts find things the way they should be. This is not because 
  FAI is faulty implemented but because "chroot" is often far too
  troublesome. 

- The use of "initrd" speeds up installation quite a bit -- in fact it
  should be sufficient to boot only once (!). 

These two features really distinguish FAI and NAIS. There are some other
significant differences (see the NAIS home page), but those can probably
be integrated into FAI, at least as "optional features". 

If you have any other questions concerning NAIS, feel free to reply using
this mailing list or send an email to "nais at informatik.uni-koeln.de". We
appreciate any suggestions, comments ore criticism.

Best regards,

Lech Nieroda, Jens Ruehmkorf, Klaus Schiwinsky, Mattias Gaertner

----------------------------------------------------------------------
               http://www.informatik.uni-koeln.de/nais/
                network automated installation system                   
----------------------------------------------------------------------