[Mono-list] GemBox.Spreadsheet 2.9 supporting Mono

[Zeljko at GemBox Software]

That is what we did eventually, we created our own method to create
pseudorandom numbers.

I need to give credit to you guys for excellent compatibility with .NET
Framework. This year we planned to add support for both Mono and .NET
Compact Framework. To our surprise, adaptation to Mono was just few days of
work, while adaptation to MS .NET Compact Framework was postponed because of
too many modification we needed to make :)

Regards,

Zeljko
GemBox Software
www.GemBoxSoftware.com 


-----Original Message-----
From: Sebastien Pouliot [mailto:sebastien.pouliot at gmail.com] 
Sent: Monday, September 10, 2007 7:39 PM
To: mono-list at lists.ximian.com; Zeljko at GemBox Software
Subject: Re: [Mono-list] GemBox.Spreadsheet 2.9 supporting Mono

My understanding is that the issue is solved but here's my two cents
anyway...

Using SHA1 is *heavyweight* (CPU) for the job of getting "regular"
randomness. A lot of faster algorithms are available to do so. E.g. You
could copy Mono's Random implementation (MIT licensed) and execute it
under any runtime (MS included).

If you need more than "regular" (which I doubt in your case) while still
keeping a replicable stream then using a HMAC (like HMAC-SHA1) would be
more suited for the task (than SHA1 alone).

Sebastien

On Mon, 2007-09-10 at 18:25 +0100, Alan McGovern wrote:
> If you need the same numbers for a given seed, why are you using a
> random number generator? If you need reproducible pseudo-random
> numbers then you could just SHA1 has your seed, which gives you 20
> bytes (4 integers). Then if you need more numbers, SHA1 hash the
> result from the previous operation to get a new set of 20 bytes (4
> integers) and keep going until you have all the numbers you need. This
> is guaranteed to produce the same psuedo-random numbers on every
> platform every time. 
> 
> Alan.
> 
> On 9/10/07, Miguel de Icaza <miguel at novell.com> wrote:
>         Hello,
>         
>         > After customer request, we have modified our .NET component
>         to work with
>         > Mono. To do so few minor changes were required and full
>         rewrite of Random
>         > methods (since .NET Framework and Mono return different
>         random values for 
>         > the same seed).
>         
>         This seems like an odd dependency to have (depending on the
>         random
>         number generator to have the same values).
>         
>         Is there a reason for that?
>         
>         Miguel
>         _______________________________________________ 
>         Mono-list maillist  -  Mono-list at lists.ximian.com
>         http://lists.ximian.com/mailman/listinfo/mono-list
> 
> _______________________________________________
> Mono-list maillist  -  Mono-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-list