Re: Evolution
From: Betov (betov_at_free.fr)
Date: 11/21/03
- Next message: Herbert Kleebauer: "Re: Evolution"
- Previous message: Randall Hyde: "Re: Writing an Adventure game with HLA"
- In reply to: Alex McDonald: "Re: Evolution"
- Next in thread: Randall Hyde: "Re: Evolution"
- Reply: Randall Hyde: "Re: Evolution"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 21 Nov 2003 22:21:57 GMT
"Alex McDonald" <alex_mcd@btopenworld.com> écrivait
news:bplnkl$gnk$1@titan.btinternet.com:
>
> "Betov" <betov@free.fr> wrote in message
> news:XnF943A5F08F47BDbetovfreefr@213.228.0.75...
>> _____________________
>> * Errors Managements:
>>
>> I am not the one who impemented a lazy SEH. Instead, RosAsm
>> Errors Management is a per cases based implementation providing
>> sensible errors messages, as well at [Compile] time as at
>> [Run] Time. So, yes (!!!...), for errors that i never thought
>> about, that i have never heard about and never seen... _yes_,
>> the _normal_ behaviour is to hang. The great thing with this
>> deliberate choice is that it _forces_ users to report, and i
>> have to say that, at least for the Assembler itself, i simply
>> don't recall the last time i had some report of that kind...
>
> Or forces users to throw it away because its broken. Exception based
> error handling is there to provide a mechanism to simplify the code
> you write; rather than have to address all the _potential_ errors
> in-line in the code (very often to the detriment of clarity),
> exception handling allows you to focus on the task in hand and treat
> exceptions as (1) infrequent(2) interrupts (3) out of line of the main
> code and (4) on a case by case basis.
>
> More importantly, it forces you to recognise that your software will
> always have a fault somewhere. Software should degrade gracefully; if
> it must fail, then with as good diagnostics as possible. Managing
> unknown errors is about recognising their inevitability; that is, _at
> design time_ making sure that they can be handled and reported at run
> time. Software that explodes, hangs, loops or dies silently is simply
> poorly designed software.
>
Might be true, generally speacking of SoftWare. For an Open
Source Auto-Compilable programming Tool, this does not make
sense to me. In hang case (hang of the RosAsm itself), user
can run a child version and report exactely where, how, and
why it hangs, with the help of the integrated Debugger.
Also, it does not make much difference, for a hanging Tool,
if it hangs silently, or if it delivers a Message with an
information quality close to zero, as the ones we usually
have with SEH.
Since all these years users have programmed with SpAsm and
RosAsm, i have never seen _one_ user message saying:
"Have this error Message: >> "....". What did i wrong?".
... Recalls you of something?
As opposed to this, i have seen, several times, users bug
reports... coming with the proper fix.
Betov.
< http://betov.free.fr/RosAsm.html >
- Next message: Herbert Kleebauer: "Re: Evolution"
- Previous message: Randall Hyde: "Re: Writing an Adventure game with HLA"
- In reply to: Alex McDonald: "Re: Evolution"
- Next in thread: Randall Hyde: "Re: Evolution"
- Reply: Randall Hyde: "Re: Evolution"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|