Re: Discovering variable types...

From: Marco van de Voort (marcov_at_stack.nl)
Date: 08/12/04 

Date: Thu, 12 Aug 2004 14:55:33 +0000 (UTC)

On 2004-08-12, J French <erewhon@nowhere.com> wrote:
> On Thu, 12 Aug 2004 09:40:26 +0000 (UTC), Marco van de Voort
><marcov@stack.nl> wrote:
><snip>
>
>>I still don't get why size is so important for you, that you are willing to
>>do so much work.
>>
>>True, decreasing complexity slightly pays back in debugging, but there is
>>a point where the total (additional development time because you are rewrite
>>everything custom per app - debugging time simple app) is still negative.
>
> Interestingly, there are similarities between the 'non mainstreamness'
> of both you and Laura
>
> You want to fit your data in memory

The future

> She wants to fit her data on a Floppy

The past

Have in common that they both different from average, but the direction
is totally different.

> You've both made good commercial cases for the original decision,

I still have good reasons.

> and in both cases I suspect that technology has invalidated those reasons.

On the contrary, the reasons are actually that technology improves (memory
is getting cheaper).

> You both have substantial code investment in the way in which it
> currently works - and in both cases it 'works' and will keep working
>
> Ironically, in both cases I've suggested disk caching, for very
> different reasons....

Which is possible if you can resell the same OLAP like read-only system over
and over again, and you can make a enormous investment in building and
optimizing such framework.

However I told you several times that the straightforwardness of the in mem
system (performance without much extra work) is the main attraction. Specially
since in this case, I can't justify making a big investment in such index
based filer system.

Relevant Pages

  • Re: xmalloc string functions
    ... a bit over three quarters of the memory, Lotus Notes, another email ... I have good reasons for using multiple email clients ... propagating the error conditions back to the caller. ... event processing against oom exceptions either. ...
    (comp.lang.c)
  • Re: xmalloc string functions
    ... I, on the other hand, can be running three VMs eating up a bit over three quarters of the memory, Lotus Notes, another email client (yes, I have good reasons for using multiple email clients simultaneously) and several other applications. ... propagating the error conditions back to the caller. ... Linux and several versions of Windows. ...
    (comp.lang.c)
  • Re: A solution for the allocation failures problem
    ... It is not possible to check EVERY malloc result within complex software. ... possible for software quality reasons. ... a memory exhaustion situation arises, ... you provoke allocation failures that don't need to happen. ...
    (comp.lang.c)
  • Re: Future Risc
    ... |> Any architecture with hw support for decimal fp is pretty CISCy indeed. ... For various reasons, I have restarted my ... simplified the code - I could do the same with the rounding, ...
    (comp.arch)
  • Re: Sun Ultra 5 not booting
    ... enough with the replacement as i ended up with 128mbs extra memory. ... But since you can't boot from CD you will need to do a network boot I guess. ... was wishing to install FreeBSD but CDrom wont boot from the disk. ... systems on SPARC hardware, receiving far less development time than on x86, and far less development time that Solaris. ...
    (comp.sys.sun.hardware)