Re: Building an assembler from scratch
- From: BGB <cr88192@xxxxxxxxxxx>
- Date: Fri, 01 Apr 2011 12:13:20 -0700
On 4/1/2011 2:18 AM, Rod Pemberton wrote:
"Rick C. Hodgin"<foxmuldrster@xxxxxxxxxxxxxxxxxxxxx> wrote in message
No problem. I'm amazed at some of the thinking in the responses.
It's like there are two completely separate trains of thought on the
matter. And I don't even understand WHY the disparity exists,
because there can be no question about it: FOSS helps everybody
all the time.
That needs clarification as to what "helps" means exactly.
"helps" reduce software development times - it depends
"helps" pay a programmer's living expenses - no.
"helps" eliminate closed source software - yes.
"helps" promote a certain political belief system - yes.
Proprietary software harms everybody most all the time,
That needs clarification as to what "harm" means exactly. I.e., just like
"help", "harm" will affect different people to different degrees.
Proprietary software benefits most people in a capitalist society.
Programmers get paid. The company pays dividends. The stock of
the company increases in value. Users get the best product, because
the company can afford to employ the best people. Those not suited
to developing the best product don't get to "destroy" it. Etc.
and since most of the world is capitalist, it is the most net benefit to follow a capitalistic model WRT software.
the prospect of being able to live off software, and thus not need to hold multiple jobs, is a possible merit of commercial development.
would most users really be benefited if their favored programs don't get updated hardly ever because all of the lead programmers are off having to work their day jobs?... does a project benefit from over-worked admins who admit crap-patches that manage to break piles of stuff, leaving someone else to try to figure out later why the newest build has suddenly got so much more buggy?...
money gives, if anything, a lot more time for developers to focus on their projects, without having to worry about all this "real life" stuff.
Proprietary software [...] only rarely does it help anybody other
than its investors / primary parties
Some proprietary software has been released as FOSS software after the
commercial lifetime of the software ended. Much proprietary software is
given away for free use already: Java, Flash, MS stuff, etc. How does that
benefit just it's investors?
Users of that software are helped, or they wouldn't pay for it. They'd see
it as a substitute good and switch products. E.g., some think OpenOffice is
as good as MS Office, and many don't.
I'll never understand how a person could _not_ choose FOSS
for their project
For one, FOSS software frequently is not up to the quality levels
expected and as provided by commercial software.
For two, FOSS, depending on the license, has some restrictions. PD doesn't.
For three, FOSS developers other than the original author, tend to destroy
the code since they frequently aren't as skilled.
For four, it's possible you could make much money, if commercial.
ANY development effort, any, that desires to obtain the maximum
possible benefit to people will ALWAYS be FOSS, period.
Commercial companies can spend millions and billions on the brightest and
best to develop the best product. Is not the "best product" the "maximum
possible benefit to people"? If not, by what criteria are you defining
"maximum possible benefit to people"?
And even further, how there can be any debate regarding the value
inherent in the GPL because it absolutely guarantees that products
released under GPL are able to have the source code, that it can
be maintained by anybody, that off-shoots can be created at any
time, etc., and forever.
As a programmer *and* a user of software, I fully understand that I do not
want source code 99.99% of the time. There is only a small percentage of
the time where I want or would like code. The vast majority of the time I
want an application that works, correctly, quickly. I want binaries. I
don't want source code that I must figure out how to compile, that is
suffering from "bit-rot", and that cannot be compiled anymore. I want
binaries. I want applications that still run 20+ years later, like those
MS-DOS. You cannot say that about Linux. You cannot even get current
version of Linux onto a mid '90's machine and have it work. They've already
started removing older hardware support from Linux. Why? Because no one
was updating the software...
backwards compatibility is IMO one of the *biggest* problem areas with Linux, likely because hardly anyone in Linux-land tends to much take it seriously...
forget binary compatibility for a moment, a lot of the software in this domain can't even really keep source-code backwards-compatibility working, given both the common practices of:
A, software using every library under the sun;
B, the maintainers of many libraries making incompatible changes to their APIs (removing API calls, changing argument lists, changing semantics and behaviors of flags and calls, ...).
at least, 15 year old source should build.
better yet is if 15 year old binaries would still work.
often, neither holds...
consider some of the recent OpenGL controversy:
OpenGL 3.2 deprecated some features (API calls);
the people maintaining the Linux drivers promptly removed them (breaking any SW which still used them, including most SW targeting the 1.x and 2.x GL standards).
meanwhile, things continue to work fine in Windows land.
all this is nevermind naive code ("glBegin()"/"glEnd()"/fixed-function pipeline/...) on my new video card and computer being not much faster than the integrated graphics on my 7 year old laptop... but it still works, this is the important part.
not everyone wants to have to continuously update/fix code to keep up with the whims of the API designers or library maintainers.
if MS were pulling this crap (well, more so than already, they have started to let things slide in recent years) they would have lost market share long ago... maybe they would have ended up, like Apple...
for example, what would Apple have now if (market dominance wise), say:
better hardware for lower costs;
backwards compatibility all the way back to the Apple II timeframe;
more "run of the mill" users would more take it seriously.
- Prev by Date: Re: Yet another handmade POPCNT
- Next by Date: Re: Yet another handmade POPCNT
- Previous by thread: Re: Building an assembler from scratch
- Next by thread: Re: Building an assembler from scratch