How to improve this sort?

[At_sea_with_C]

Hello all,

I have written an ascending sort routine for floats. This seems to do
the job, but for elements over 10,000, it gets awfully slow. A lot of
useless comparisions with previously sorted elements are made. I was
thinking of using a function, that kept an index, to selectively iterate
over unsorted elements. But I am unable to think of a way to do it.

Any ideas to improve the routine and get a better grade?
Thanks to all.

#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>

#define MAX_ALLOC 1000000UL

void sortf_asc(float *a, unsigned long elem) {
int sort_occured = 0;
unsigned long start_offset = 0, current_pos, c1;
float tmp;
while(start_offset < elem - 2) {
for(c1 = 1 + (current_pos = start_offset); c1 < elem; c1++) {
if(a[current_pos] a[c1]) {
tmp = a[c1];
a[c1] = a[current_pos];
a[current_pos] = tmp;
current_pos = c1;
sort_occured = 1;
}
}
if(sort_occured) sort_occured = 0;
else start_offset++;
}
return;
}

int main(int argc, char **argv) {
char *usage = "Usage:\n\nPROGRAM ELEMENTS\n\nELEMENTS - Number of "
"elements to sort, (max. = %lu)\n",
*fopenfail = "Failed to open file.\n",
*tail = NULL;
unsigned long elements = 0, c1;
float *farray = NULL;
FILE *outfp = NULL;

if(argc < 2){ fprintf(stderr, usage, MAX_ALLOC); return EXIT_FAILURE; }
else {
errno = 0;
elements = strtoul(argv[1], &tail, 0);
if(errno == EINVAL || errno == ERANGE) {
fprintf(stderr, usage, MAX_ALLOC);
return EXIT_FAILURE;
}
else if(elements < 2 || elements MAX_ALLOC) {
fprintf(stderr, usage, MAX_ALLOC);
return EXIT_FAILURE;
}
}

farray = malloc(elements * sizeof *farray);
if(!farray) {
fprintf(stderr, "Memory allocation failed.\n");
return EXIT_FAILURE;
}
else {
srand((unsigned int)time(NULL));
for(c1 = 0; c1 < elements; c1++) farray[c1] = (float)rand();
}

outfp = fopen("unsorted", "w");
if(!outfp) {
fprintf(stderr, fopenfail);
free(farray);
return EXIT_FAILURE;
}
else {
for(c1 = 0; c1 < elements; c1++)
fprintf(outfp, "[%lu]: %f\n", c1, farray[c1]);
fclose(outfp);
outfp = NULL;
}

sortf_asc(farray, elements);

outfp = fopen("sorted", "w");
if(!outfp) {
fprintf(stderr, fopenfail);
free(farray);
return EXIT_FAILURE;
}
else {
for(c1 = 0; c1 < elements; c1++)
fprintf(outfp, "[%lu]: %f\n", c1, farray[c1]);
fclose(outfp);
}

return 0;
}
--
Email: The handle, (dot seperated), at gmail dot com.

[Christopher Layne]

Chris Dollin wrote:

>What different does it make for simple replies? Also, do you not keep a
temporary archive of last posts in your newsreader?

I don't. Why should I have to?

Likewise.

[Christopher Layne]

CBFalconer wrote:

For one thing, the attribution allows one to evaluate the validity
of quoted material (in some cases). For example, if it comes from
werty, MI5, or Kenny McCormack it is immediately know to be
worthless. If it comes from Ben Pfaff, Chris Torek, Richard
Heathfield, it is expected to be pertinent. Just sample names that
immediately come to mind, omission (from either list) is not
significant. Inclusion is.

So are you admitting that you establish bias before even reading the quoted
material? Quite modern.

[Richard Bos]

Christopher Layne <cl****@com.anodizedwrote:

CBFalconer wrote:

For one thing, the attribution allows one to evaluate the validity
of quoted material (in some cases). For example, if it comes from
werty, MI5, or Kenny McCormack it is immediately know to be
worthless. If it comes from Ben Pfaff, Chris Torek, Richard
Heathfield, it is expected to be pertinent. Just sample names that
immediately come to mind, omission (from either list) is not
significant. Inclusion is.

So are you admitting that you establish bias before even reading the quoted
material? Quite modern.

Everybody does this. Not everybody admits to it. Perhaps being honest is
quite modern where you live; I have to say that over this way it's
considered an old-fashioned virtue.

Richard

[Richard Heathfield]

Christopher Layne said:

CBFalconer wrote:

>For one thing, the attribution allows one to evaluate the validity
of quoted material (in some cases). For example, if it comes from
werty, MI5, or Kenny McCormack it is immediately know to be
worthless. If it comes from Ben Pfaff, Chris Torek, Richard
Heathfield, it is expected to be pertinent. Just sample names that
immediately come to mind, omission (from either list) is not
significant. Inclusion is.

So are you admitting that you establish bias before even reading the
quoted material?

Of course. I do it every time I walk into a bookshop. It's full of quoted
material, but one glance at the attribution on the cover is normally
sufficient to give me the information I need on whether the quoted material
is something I want to spend my time investigating. For example, if it's
Kernighan, Stewart, or Pratchett, the answer is an unequivocal "yes". If
it's Schildt, the answer is an unequivocal "no" (except perhaps out of a
kind of macabre fascination with the unspeakably evil).

Quite modern.

Kind of you to say so, but being discerning in my reading choice is
something I've been doing for decades.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Richard Heathfield]

CBFalconer said:

Richard Tobin wrote:

>user923005 <dc*****@connx.comwrote:

>>>Do you mean existing library implementations? Allowing the
comparison function in qsort() to call qsort() couldn't break
any user code.

Who wrote the above?

Richard Tobin.

>>If qsort() calls compare() and compare() calls qsort(), then it
will break things if qsort() is not reentrant.

Yes, exactly. I'm wondering why the standard didn't require this
to work. Presumably the answer is just that some existing qsort()
implementations (in the 1980s) weren't re-entrant.

Why do you want to hide the author of the first quote? That is, to
my mind, unethical. In fact it could be considered plagiarism.

Plagiarising one's own work is not normally considered unethical.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Richard Heathfield]

Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a few. The
optimum level is <= N, where N depends on what floats your boat. For me,
N = 6 (for various values of 6).

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Default User]

Richard Heathfield wrote:

Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a few.
The optimum level is <= N, where N depends on what floats your boat.
For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted. Once
you get that deep into quote levels, the chances are pretty good that
the earlier stuff needs to be excised completely. If it is still
relevant, then the attributions are as well.


Brian

[Christopher Layne]

Richard Heathfield wrote:

>So are you admitting that you establish bias before even reading the
quoted material?

Of course. I do it every time I walk into a bookshop. It's full of quoted
material, but one glance at the attribution on the cover is normally
sufficient to give me the information I need on whether the quoted material
is something I want to spend my time investigating. For example, if it's
Kernighan, Stewart, or Pratchett, the answer is an unequivocal "yes". If
it's Schildt, the answer is an unequivocal "no" (except perhaps out of a
kind of macabre fascination with the unspeakably evil).

Yes - I do this as well. You know why? Because I most likely have to open the
book and wade through it.

But a paragraph of contextual information like a usenet reply usually doesn't
require as much pre-measurement to just read and then determine
afterwards "looks like crap or not." At that people you can then check the
sender and see, oh Richard Heathfield wrote that - there's probably some
validity to it and maybe I should examine why I thought it looked bogus in
the first place.

[Keith Thompson]

Christopher Layne <cl****@com.anodizedwrites:
[...]

But a paragraph of contextual information like a usenet reply usually doesn't
require as much pre-measurement to just read and then determine
afterwards "looks like crap or not." At that people you can then check the
sender and see, oh Richard Heathfield wrote that - there's probably some
validity to it and maybe I should examine why I thought it looked bogus in
the first place.

Most newsreaders (in fact, all the ones I know of) provide attribution
lines automatically. Even Google Groups does so. All you have to do
is leave them in place rather than manually deleting them.

Quoting someone else's words without attribution is rude, especially
when it's easier to provide attribution than to delete it.

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.

[Richard Heathfield]

Christopher Layne said:

<snip>

But a paragraph of contextual information like a usenet reply usually
doesn't require as much pre-measurement to just read and then determine
afterwards "looks like crap or not." At that people you can then check the
sender and see, oh Richard Heathfield wrote that - there's probably some
validity to it and maybe I should examine why I thought it looked bogus in
the first place.

This happened to me in comp.programming recently. Some guy posted what
looked at first glance like a load of bizarre techno-babble about some new
data structure or other. "Yeah, right", thought I. "Who wrote this junk
anyway?" Well, it turned out to be Ben Pfaff. Which, of course, meant that
it wasn't junk at all, because Ben Pfaff doesn't waste time with junk - and
that meant I had to read the whole darn thing over again to find out what
he was actually getting at.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Richard Heathfield]

Default User said:

Richard Heathfield wrote:

>Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a few.
The optimum level is <= N, where N depends on what floats your boat.
For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted.

In my case, not, because I normally work pretty hard to remove unnecessayr
text. But sometimes I feel it wise to leave in several levels of quoting.
In the case where it's A and B ping-ponging:

A said:
| B said:
|| A said:
||| B said:
|||| A said:
||||| B said:

....I don't think it terribly unreasonable to remove the innermost attribs.

(Normally, of course, I leave attribs well alone.)

<snip>

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Default User]

Richard Heathfield wrote:

Default User said:

Richard Heathfield wrote:

Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a few.
The optimum level is <= N, where N depends on what floats your

boat. >For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted.

In my case, not, because I normally work pretty hard to remove
unnecessayr text. But sometimes I feel it wise to leave in several
levels of quoting. In the case where it's A and B ping-ponging:

A said:
| B said:
|| A said:
||| B said:
|||| A said:
||||| B said:

...I don't think it terribly unreasonable to remove the innermost
attribs.

(Normally, of course, I leave attribs well alone.)

It's unlikely that all that quoted stuff is needed. Once it gets past
about four levels, I tend to just paraphrase the previous discussion.

If I respond to a ping-pong message, it's usually to whinge at the
participants for not snipping.


Brian

[Christopher Layne]

Most newsreaders (in fact, all the ones I know of) provide attribution

lines automatically. Even Google Groups does so. All you have to do
is leave them in place rather than manually deleting them.

Quoting someone else's words without attribution is rude, especially
when it's easier to provide attribution than to delete it.

There are times when the discussion is so deeply drawn out and localized that
the scope is well established and clipping a few lines here or there to
reduce cruft in the reply (for readiblity) does not affect anyone - as they
may not even be addressing the issue of *who* wrote something.

For instance, I clipped your attribution here, just to have less lines to
read/send out. Did it change anything at all about how you read it? I
seriously doubt it - as the scope is now fairly local and our memory of it is
not so short term as to be cleared within 5 seconds.

That being said I usually include them - but I don't *flip out* if they're
not. I have more issue with whiners in this newsgroup not being able to
*adapt* and choosing everything as a battle rather than choosing wise
battles.

How many times must Chuck lecture people? I get tired of reading the
admonishments and I get tired of the waste of bandwidth. When I see it, I end
up <space>-ing right over it because I immedaitely expect the reply to just
be a lecture. I personally am not interested in reading lectures in
comp.lang.c - I'd rather read about constructive concepts and posts which
teach myself and others new things.

Why not just disengage from the top posters and whatever other devil-child
bothers you at that moment? The fact that someone is going out of their way
to type up *another* post just to tell that other person how to post
(especially to original posts which didn't have much content nor thinking on
their own right) is pretty pavlonian.

For instance, foreign language speaker arrives in group and posts a cryptic 2
line post asking how to do something w/ an additional 5 lines of
implementation specific crappy code - complete with bad style, grammar,
and "requirements". Standard reaction: People go after it and start
instructing the person on proper grammar, proper usage of the standard usage
of main(), etc. etc. etc.

Is it so hard to just ignore the obviously not thought out lazy posts from
people who just want an answer?

[Keith Thompson]

Christopher Layne <cl****@com.anodizedwrites:

Most newsreaders (in fact, all the ones I know of) provide attribution
lines automatically. Even Google Groups does so. All you have to do
is leave them in place rather than manually deleting them.

Quoting someone else's words without attribution is rude, especially
when it's easier to provide attribution than to delete it.

I wrote the above. You deliberately deleted the attribution line.

[...]

For instance, I clipped your attribution here, just to have less lines to
read/send out. Did it change anything at all about how you read it? I
seriously doubt it - as the scope is now fairly local and our memory of it is
not so short term as to be cleared within 5 seconds.

Ok, so you snipped the attribution line to make a point. I'm not
impressed. It's still rude.

This is a busy newsgroup. Some of us, myself included, read
everything posted here, or nearly so. When I read a followup, I might
have read a dozen other threads since I read the original article.
Any little cue that can remind me of the context is helpful.

[...]

Why not just disengage from the top posters and whatever other devil-child
bothers you at that moment?

[...]

Because some people are capable of learning. People who initially
top-post, when told that it's not generally accepted here, sometimes
actually *stop top-posting* and become useful members of the
newsgroup. People who use silly abbreviations like "u" and "ur"
sometimes drop them and write in standard English. There are posters
I consistently ignore; I do not choose to add someone to that list for
a single infraction.

If you don't like what I post here, why not add me to your killfile?
If you don't like what CBFalconer posts here, why not add him to your
killfile?

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.

[CBFalconer]

Richard Heathfield wrote:

Default User said:

>Richard Heathfield wrote:

>>Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a
few. The optimum level is <= N, where N depends on what floats
your boat. For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted.

In my case, not, because I normally work pretty hard to remove
unnecessayr text. But sometimes I feel it wise to leave in several
levels of quoting. ...

With properly posted text it is usually simple. Snip as required,
then count the '>'s of the uppermost remaining quotation, and snip
all attributions with that much (or more) '>'s. You may have to
make allowance for idiots who have perverted the quote marker, but
I find the human mind is quite capable of that adjustment. You
also have to adjust for other idiots who post overly long lines, so
that the wraps get marked with insufficient markers.

--
<http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>

"A man who is right every time is not likely to do very much."
-- Francis Crick, co-discover of DNA
"There is nothing more amazing than stupidity in action."
-- Thomas Matthews

[CBFalconer]

Christopher Layne wrote:

>

.... snip ...

>
How many times must Chuck lecture people? I get tired of reading
the admonishments and I get tired of the waste of bandwidth. When
I see it, I end up <space>-ing right over it because I immedaitely
expect the reply to just be a lecture. I personally am not
interested in reading lectures in comp.lang.c - I'd rather read
about constructive concepts and posts which teach myself and
others new things.

Then why don't you simply conform to standard practice and avoid
being lectured? Why do you rudely snip attributions, and thus
steal the words of others? This effectively amounts to plagiarism.

.... snip ...

>
For instance, foreign language speaker arrives in group and posts
a cryptic 2 line post asking how to do something w/ an additional
5 lines of implementation specific crappy code - complete with bad
style, grammar, and "requirements". Standard reaction: People go
after it and start instructing the person on proper grammar,
proper usage of the standard usage of main(), etc. etc. etc.

Those posters are normally simply ignorant of proper practices, and
an early and immediate advisement will often turn them into good
net citizens. Then there are the others, who persist in rejecting
normal practice and thus engendering many attempts at correction.
In children, this kind is usually considered a bad child, and the
neighbors children are forbidden to play with him/her. The usenet
analog is PLONK.

I am quite happy to forgo any corrections, if the corrections were
not needed or have already been made. Someone said "Vigilance is
the price of freedom". Until recently this country seemed to have
forgotten that, while allowing the shrub to emasculate rights in
favor of his adventure. We hope to avoid an analagous interregnum
of chaos on this newsgroup.

--
<http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>

"A man who is right every time is not likely to do very much."
-- Francis Crick, co-discover of DNA
"There is nothing more amazing than stupidity in action."
-- Thomas Matthews

[Ben Pfaff]

CBFalconer <cb********@yahoo.comwrites:

Why do you rudely snip attributions, and thus steal the words
of others? This effectively amounts to plagiarism.

Plagiarism is presenting another's work as your own. Quoting
without attribution is not plagiarism, because it does not claim
the quoted words as one's one. It is merely use without complete
citation.
--
Ben Pfaff
bl*@cs.stanford.edu
http://benpfaff.org

[Default User]

Christopher Layne wrote:

Most newsreaders (in fact, all the ones I know of) provide
attribution lines automatically. Even Google Groups does so. All
you have to do is leave them in place rather than manually deleting
them.

Quoting someone else's words without attribution is rude, especially
when it's easier to provide attribution than to delete it.

There are times when the discussion is so deeply drawn out and
localized that the scope is well established and clipping a few lines
here or there to reduce cruft in the reply (for readiblity) does not
affect anyone - as they may not even be addressing the issue of who
wrote something.

Removing the attribution doesn't do any of that, in fact it makes the
entire post less readable.


Brian

[Richard Heathfield]

CBFalconer said:

<snip>

I am quite happy to forgo any corrections, if the corrections were
not needed or have already been made. Someone said "Vigilance is
the price of freedom". Until recently this country seemed to have
forgotten that,

Chuck - this is an international group. There is no such thing as "this
country", without some kind of qualification. As far as I'm concerned, an
unqualified "this country" means the United Kingdom of Great Britain and
Northern Ireland, but your mileage may vary, and so might other people's.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Richard Heathfield]

CBFalconer said:

Richard Heathfield wrote:

>Default User said:

>>Richard Heathfield wrote:
Default User said:

<snip>

Why is it worthwhile deleting [attributions]?

The more there are, the more worthwhile it becomes to remove a
few. The optimum level is <= N, where N depends on what floats
your boat. For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted.

In my case, not, because I normally work pretty hard to remove
unnecessayr text. But sometimes I feel it wise to leave in several
levels of quoting. ...

With properly posted text it is usually simple.

Yes, Chuck, I *know*, Chuck - and I *usually* leave the attributions in,
Chuck, because *usually* it's the right thing to do, with *usually* being
the operative word. Which bit of "sometimes" were you struggling with?

Snip as required,
then count the '>'s of the uppermost remaining quotation, and snip
all attributions with that much (or more) '>'s.

Cf <a5********************@bt.com>

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[CBFalconer]

Richard Heathfield wrote:

CBFalconer said:

>Richard Heathfield wrote:

>>Default User said:
Richard Heathfield wrote:
Default User said:
>
<snip>
>
>Why is it worthwhile deleting [attributions]?
>
The more there are, the more worthwhile it becomes to remove a
few. The optimum level is <= N, where N depends on what floats
your boat. For me, N = 6 (for various values of 6).

That generally means that unnecessary text has been left quoted.

In my case, not, because I normally work pretty hard to remove
unnecessayr text. But sometimes I feel it wise to leave in several
levels of quoting. ...

With properly posted text it is usually simple.

Yes, Chuck, I *know*, Chuck - and I *usually* leave the attributions
in, Chuck, because *usually* it's the right thing to do, with
*usually* being the operative word. Which bit of "sometimes" were
you struggling with?

I'm not struggling. You, however, seem to be confusing advice
aimed at the general readership with criticism of your post.
Rather, it was amplification, and by default should be so
considered.

>

>Snip as required,
then count the '>'s of the uppermost remaining quotation, and snip
all attributions with that much (or more) '>'s.

--
<http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>

"A man who is right every time is not likely to do very much."
-- Francis Crick, co-discover of DNA
"There is nothing more amazing than stupidity in action."
-- Thomas Matthews

[Richard Heathfield]

CBFalconer said:

Richard Heathfield wrote:

>>
Which bit of "sometimes" were you struggling with?

I'm not struggling. You, however, seem to be confusing advice
aimed at the general readership with criticism of your post.
Rather, it was amplification, and by default should be so
considered.

Must. Not. Post. Before. Coffee.

Sorry, Chuck.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Christopher Layne]

Default User wrote:

Removing the attribution doesn't do any of that, in fact it makes the
entire post less readable.

Brian

Try reading the damn post first dude - not who posted it. If you need that
information, you can find it. It doesn't make the post ANY less readable.

[Christopher Layne]

CBFalconer wrote:

Then why don't you simply conform to standard practice and avoid
being lectured? Why do you rudely snip attributions, and thus
steal the words of others? This effectively amounts to plagiarism.

I'm sorry, the quote indicators are there. It isn't plagarism. Cut the
melodrama.

Those posters are normally simply ignorant of proper practices, and
an early and immediate advisement will often turn them into good
net citizens. Then there are the others, who persist in rejecting
normal practice and thus engendering many attempts at correction.
In children, this kind is usually considered a bad child, and the
neighbors children are forbidden to play with him/her. The usenet
analog is PLONK.

Bad child? Neighbors kids forbidden to play with him/her? We're adults here.
This is a usenet group for discussion of the C language - not a supreme
court.

I am quite happy to forgo any corrections, if the corrections were
not needed or have already been made. Someone said "Vigilance is
the price of freedom". Until recently this country seemed to have
forgotten that, while allowing the shrub to emasculate rights in
favor of his adventure. We hope to avoid an analagous interregnum
of chaos on this newsgroup.

Seek help.

[Richard Heathfield]

Christopher Layne said:

This is a usenet group for discussion of the C language - not a
supreme court.

Then could you please stop trying to bash people over the head and start
discussing C instead? Thank you.

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.

[Default User]

Christopher Layne wrote:

Default User wrote:

Removing the attribution doesn't do any of that, in fact it makes
the entire post less readable.

Brian

Try reading the damn post first dude - not who posted it. If you need
that information, you can find it. It doesn't make the post ANY less
readable.

I do not go searching for information. The post should contain all the
context it requires, that includes attributions. And in spite of your
claim, yes it does make it less readable. Not in the "follow the words"
sense, but in figuring out what's being said by whom.

You still haven't given any logical reason for removing attributions.
Note, it's exactly that. Every newsreader I know of automatically puts
them in, you have make an explicit effort to take them out.

Brian