Re: network transparency (Re: GNOME CVS: gnome-core mmclouglin)
- From: Mathieu Lacage <mathieu gnu org>
- To: Alan Cox <alan redhat com>
- Cc: Havoc Pennington <hp redhat com>, Rodrigo Moya <rodrigo gnome-db org>, Jeff Waugh <jdub perkypants org>, GNOME Hackers <gnome-hackers gnome org>, GNOME Components <gnome-components-list gnome org>
- Subject: Re: network transparency (Re: GNOME CVS: gnome-core mmclouglin)
- Date: 03 Dec 2001 09:26:49 +0100
le dim 02-12-2001 à 20:02, Alan Cox a écrit :
> > networking; i.e. nice interfaces for doing networking, that go ahead
> > and expose the things you need to worry about, such as asynchronicity
> > and latency, and give you nice tools to manage them. Burying those
> > things under "transparency" is not a win, it just means people don't
> > deal with them.
>
> Actually there are two approaches, and there are good arguments for either
>
> 1. You don't pretend and you expose the reality to apps.
> [The "core doesnt have enough info to anticipate" argument]
>
> 2. You embed it into the language itself
> [The "only the core can see the whole picture and make global
> decisions" argument]
>
> It's possible to take a smalltalk or squeak application and make the entire
> thing distributed with close to no changes to the application for example.
This is possible in Ada95 too but making an application distributed by
adding the right pragmas here and there will not make it "cool" all of a
sudden. Unless it was designed and componentized specifically for the
distributed case, it will work but suck.
Mathieu
>
> Alan
--
Mathieu Lacage <mathieu gnu org>
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]