Darren New <dne### [at] sanrrcom> wrote:
> Oh, and Warp, they aren't "unsigned integers" and "signed integers" in 
> C. They're "ints", not "integers".  ;-) That's the root of the problem. 
> Mixing signed integers and unsigned integers are fine. It's just C's 
> weird conversion rules for ints that are problematic. ;-)  </nit>

  So the implicit conversion rules of C for signed and unsigned integers
are weird? Care to suggest better conversion rules?

-- 
                                                          - Warp

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> There's the problem right there.  As one professor I knew put it, "There 
> are two kinds of languages. The ones that support unbounded arithmetic 
> types, and the ones that don't know they need to support unbounded 
> arithmetic types."

  Having automatic unbounded arithmetic types is fine as long as you don't
care about efficiency.

  Integral types which support unlimited precision can *not* be as efficient
as CPU-register-sized integers unless you explicitly tell to the compiler
that "this variable will always stay within these limits", in which case
you are already stuck with the same limitation as the C integral variables.
If you don't tell the compiler those limits and it has to make sure that
in case of overflow it switches to unlimited precision, then those integral
types simply cannot be as fast as the bounded ones. It's just physically
impossible. If nothing else, the compiler will have to add an overflow
check after each single operation done with those integers, thus adding
clock cycles and code size (filling code caches faster).

  I'm also sure that being forced to prepare for unlimited precision math
makes many compiler optimizations impossible (which would be possible with
register-sized integers).

  (Also, in the general case a compiler cannot deduce by examining a piece
of code that a variable will never have a value outside certain boundaries.
I'm certain this kind of check would be equivalent to the halting problem.)

-- 
                                                          - Warp

Post a reply to this message

Aw quit yer complainin' and write your own microscopic software if you 
wants it.  That's the proper *nix-user attitude.  XD

-- 
Tim Cook
http://home.bellsouth.net/p/PWP-empyrean

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GFA dpu- s: a?-- C++(++++) U P? L E--- W++(+++)>$
N++ o? K- w(+) O? M-(--) V? PS+(+++) PE(--) Y(--)
PGP-(--) t* 5++>+++++ X+ R* tv+ b++(+++) DI
D++(---) G(++) e*>++ h+ !r--- !y--
------END GEEK CODE BLOCK------

Post a reply to this message

Tim Cook <z99### [at] bellsouthnet> wrote:
> Aw quit yer complainin' and write your own microscopic software if you 
> wants it.  That's the proper *nix-user attitude.  XD

  That's right. Open Office is open source, so it's a good start. Go and
optimize it until it's small enough for your taste.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Tim Cook <z99### [at] bellsouthnet> wrote:
>> Aw quit yer complainin' and write your own microscopic software if you 
>> wants it.  That's the proper *nix-user attitude.  XD
> 
>   That's right. Open Office is open source, so it's a good start. Go and
> optimize it until it's small enough for your taste.

Arguably it would be faster to start again from scratch rather than try 
and learn the OO codebase... :-S

Post a reply to this message

On Wed, 24 Oct 2007 18:57:07 +0100, Orchid XP v7 wrote:

> Indeed no - their job is to research new techniques for slowing software
> down as much as possible to boost sales of expensive new hardware.
> (Presumably this is why the hardware vendors love them so much...)

That's been my working theory for years.  Goes something like this:

NetWare:  Solid, stable, reliable OS.  No possibility for hardware 
upgrades (until 4.x came out) or extended support contracts, because it 
just worked.

Windows:  Higher hardware requirements, not as stable as NetWare, so 
requires ongoing maintenance, support, and you hope that they'll fix the 
bugs in the current version in the next version so you have an upgrade 
opportunity.  Of course, with the new version comes higher hardware 
requirements, so the vendor gets to make their hardware sales quotas, and 
you'll want support as well, so a reseller can also sell a support 
contract with the hardware and SW installation.

Things have changed a bit over the last few years, of course, but I never 
wondered why resellers didn't push NetWare over Windows, and I still 
don't.  There's no money in it for them.

I've always said that Microsoft was the king of making software that was 
"just good enough".  IOW, it's good enough that most people are not 
willing to go to the expense of switching, even though there's plenty to 
complain about.  They then managed to find the right upgrade price point 
that played to the business people - the point at which it's too 
expensive *immediately* to switch to something better as opposed to just 
getting the upgrades and praying.  Businesses that tend to look more than 
a quarter or two out on their financial planning tend to be more likely 
to switch; businesses where turnover is fairly high and people are only 
concerned about the next couple of quarters' worth of spending are less 
likely to switch.

Jim

Post a reply to this message

On Wed, 24 Oct 2007 11:26:11 -0700, Darren New wrote:

> Jim Henderson wrote:
>> To someone with a little background in C
> 
> There's the problem right there.  As one professor I knew put it, "There
> are two kinds of languages. The ones that support unbounded arithmetic
> types, and the ones that don't know they need to support unbounded
> arithmetic types."

Yep, I can see that - the need for arbitrary-size variables (and the 
supporting math libraries) has become much more important.

That said, though, I'm talking about a product developed over 10 years 
ago, and from having seen bits of the source code and knowing many of the 
developers, that would've been the "language of choice" at the time.

> I can't tell you how many tools I went through looking for one that
> would handle a file >4G for restartable downloads. I finally had to
> install bittorrent at both ends just to move the file across the
> network.  In spite of the fact that FTP uses ASCII representations for
> all numbers, and hence has no inherent limitation in the protocol.

I see that problem all the time myself as well - I make downloads 
available for a limited audience using a tool that has problems with IE7 
(of all things) because they're too big - FF doesn't have a problem with 
them, and many of the download managers we suggest also work fine (I use 
wget and never have a problem with > 4 GB file sizes).

Jim

Post a reply to this message

On Wed, 24 Oct 2007 20:22:30 +0100, Orchid XP v7 wrote:

> Arguably it would be faster to start again from scratch rather than try
> and learn the OO codebase... :-S

Doubt it, otherwise new coders coming into the project would never get 
started (into any project, for that matter).

I got access to the code for one of the products I used to support, and 
it was several million lines of code.  The engineering manager who 
granted me the access said they *normally* spend at least 6 solid months 
of examining and learning the code before they're allowed to commit 
changes.

And that's a minimum of 8 hours a day of doing code review...for 6 
months...

Jim

Post a reply to this message

Jim Henderson wrote:
> On Wed, 24 Oct 2007 20:22:30 +0100, Orchid XP v7 wrote:
> 
>> Arguably it would be faster to start again from scratch rather than try
>> and learn the OO codebase... :-S
> 
> Doubt it, otherwise new coders coming into the project would never get 
> started (into any project, for that matter).

I always wondered about this... How *do* you get started on a new 
project with millions of lines of code and no documentation?

(I've looked at a couple of such projects, but failed to learn anything 
of use. I just couldn't figure out what the heck any of it is...)

Post a reply to this message

"Orchid XP v7" <voi### [at] devnull> wrote in message
news:471fa7b9$1@news.povray.org...
> Jim Henderson wrote:
> > On Wed, 24 Oct 2007 20:22:30 +0100, Orchid XP v7 wrote:
> >
> >> Arguably it would be faster to start again from scratch rather than try
> >> and learn the OO codebase... :-S
> >
> > Doubt it, otherwise new coders coming into the project would never get
> > started (into any project, for that matter).
>
> I always wondered about this... How *do* you get started on a new
> project with millions of lines of code and no documentation?

Usually with the help of someone who knows the codebase and quite a bit of
time. I've done it a couple times. Doesn't really get easier.

Post a reply to this message