Linux V embedded MS. I need a paper that refutes this:

http://www.embedded-forecast.com/EMFTCD2003v3.pdf

Someone quoting the report is not good enough, I need solid figures as
per the origonal report, preferably supporting linux.

This is by no means meant to start linux v MS flamewar, just providing
management with costing info and direction on a new product.

The Real Andy wrote:
> http://www.embedded-forecast.com/EMFTCD2003v3.pdf
> 

Either this is total bullshit, or we are exceptionally productive here.

Our recent Embedded Linux design:
  - time to market: 1 year
  - software engineers: 2 (for both Linux programming and VHDL coding)
  - no prior embedded Linux experience, nor experience with the chosen 
CPU (Xscale)
  - software engineering Man Months related to Linux coding: 12 (roughly)

Regards,
Andrzej Ekiert
http://andrzejekiert.ovh.org/

Andrzej Ekiert wrote:
> The Real Andy wrote:
> > http://www.embedded-forecast.com/EMFTCD2003v3.pdf
> >
>
> Either this is total bullshit, or we are exceptionally productive here.
>
> Our recent Embedded Linux design:
>   - time to market: 1 year
>   - software engineers: 2 (for both Linux programming and VHDL coding)
>   - no prior embedded Linux experience, nor experience with the chosen
> CPU (Xscale)
>   - software engineering Man Months related to Linux coding: 12 (roughly)
>
> Regards,
> Andrzej Ekiert
> http://andrzejekiert.ovh.org/

Total Cost of Development is based on "# of Average Software
Engineers."

Exceptional software engineers don't enter into it. ;^)

Not a refutal, but an interesting read about this paper:
<http://www.linuxdevices.com/articles/AT2156107754.html>

This reminds me about one popular Linux vs/ Windows server benchmark.
Server "A", was saturating the Ethernet link, but server "B" was
"better" because it transferred data (over the network) twice as fast...