touche

Has anyone else done the optimizing tests on 2012? If there is a new pecking order, how does that affect you?

Is G++ a runner in this race at all?

that was supposed to be a jab at C++, not you (or only you in as much as you subject yourself to the tortures of C++ voluntarily)

Can you just swap the intel compiler for another and expect to get better results or is there some tuning that needs to be done? Intel 13 seemed to under perform gcc when I tried it, however I only had time to basically compile it and try swapping it for gcc (was just looking for an easy optimisation and will be exploring the issue further when I have more time).

General programming question:

Do you guys think programming is the art of problem solving? I've seen people align the two closely, and it seems many programmers just like to solve problems. For example, many like solving arbitrary (if I may use the word) math problems (such as Project Euler). I've never really found myself a problem solver like that, and usually think more than do or "tinker".

So, is programming about problem solving?

It depends. The correlation you noticed is true, though I wouldn't say it's necessarily true. Writing good, clean code is in many ways akin to writing good, clear prose. It's about communication, which requires understanding how others think and how to make ideas clear to them. I think someone who excelled at just this could become very good at certain kinds of programming (most web development, eg).

Of course, in practice, many types of programming are intertwined with problem solving, and you want a person who is able to both solve the problem and implement the solution in code. Strictly speaking, the best person to do the former will rarely be the best person to do the latter imo (in much the same way that a scholar might be able to explain the ideas of a philosopher more clearly than the philosopher's own works do). In practice though, people who can solve difficult problems are at least good enough programmers that only one person is hired to do both jobs. And the reverse is not true.

Has anyone ever went to a programmer's meetup? I've been thinking about going to a Clojure meetup but I keep on missing the dates. Tonight's meeting is in Cascalog, a Prolog-like language that connects with Hadoop.

From what I've seen, it appears these talks and demos are like "Look at this cool **** I created." I guess the other part is networking and making friends.

Has anyone ever been to one of these gatherings? What is the etiquette? Am I still too new to go to these things? The talks have been pretty smart from what I've seen and I fear I'll generally glaze over, ask a really stupid question, and then I could see someone asking what I thought of the talk and betraying my stupidity.

dave,

i have been to many meetups. they range in quality from boring to pretty great.

your characterization is reasonable, though i'd say it's about half look at my cool **** and half here is a concept or tool or practice that i find interesting. there's usually announcements of people who are hiring, which is like a more direct form of networking i suppose .

the etiquette is you show up, grab a beer and some pizza, and mingle for a while. if you want to hide from social interaction, bring your laptop and hack on stuff.

you can wait for a topic that particularly interests you, but i think your passion for clojure is enough of a common bond with other clojure enthusiasts that you will enjoy what they enjoy. really the important thing is to go to one and see what the group is like.

you are definitely not too new. sure, some talks will be over your head, but you'll usually learn at least a little bit by trying. you'll need to get over worrying about "looking stupid" though. experienced programmers will discern your skill level very quickly. approach your noobdom with humility and honesty, as i've seen you do countless times in this forum, and you'll get more explanations than you can handle.

programming is a discipline which is very difficult to learn in a vacuum. if we didn't have a strong tradition of the transfer of knowledge from guru to neophyte, we'd never get anything done. we've all been there and i'm sure this group will be thrilled to see new blood.

tl;dr: get off your ass and go meet some clojure heads!

Thanks tyler. I'll attend the next one no matter what it is. I was sort of looking forward to this since Prolog -> Lisp is something I would understand and enjoy. The Hadoop stuff, though...

Humility: I try, but I've definitely had my spectacular failures.

***

C is a totally awesome language. Why is C++ so reviled?

spend 5 minutes reading about what hadoop is and you're well-armed to hear a talk about it.

c is a fine language, but there's plenty to hate about it. c++ amplifies a lot of those bad qualities and also bolts on a ton of extra stuff.

tl;dr: it's a way to do parallel processing via scatter-gather. templates.

I dunno how accurate this stuff is, but I stumbled upon it a long time ago and found it an entertaining read: http://yosefk.com/c++fqa/

I have no desire to read all of that, but wow. Talk about something birthed from hatred.

I was thinking someone would simply link to http://harmful.cat-v.org/software/c++/linus and be done with it.

C++ was written for its time, and that imposed constraints.

C++ also has certain libraries that were added over time in ways that to some, seem a bit silly. But you don't have to use those, so don't let stuff like templates deter you.

If you want to use C++ in practice, for real stuff, Qt is a good library to look at in place of something like STL.

I should have mentioned that he has a pretty good tl;dr version:
http://yosefk.com/c++fqa/defective.html
http://yosefk.com/c++fqa/picture.html

There was a "ask me-thread" on the finance and investing forum here. Unfortunately I can not remember the name of the poster or name of the thread but must had been last year (at least the latest reply was). He covered pretty well different aspects of banner advertisement, design , what works etc

Not sure if this helps but won't hurt to read it. Rather interesting stuff

found it

http://forumserver.twoplustwo.com/30...-amaa-1208283/

Really like this explanation, and think it is interesting how you point out that a scholar can better explain a philosopher's theory than the philosopher himself.

fwiw: On my own journey of self-discovery, and highly self-aware (possibly to the point of weighing small details too much). Deciding the degree to which I believe CS will fit me, and think maybe I have not had complete exposure to "problem solving" in the world of CS as my problem solving has been guided through courseware.

heh... I see. I recognize that what I wrote can certainly read a little emo... so either way, I can take it


the intel compilers integrate really cleanly with visual studio... you can switch back and forth with literally a single click. Its mostly for this reason that I've been comparing between only these two recently.

some years ago, I did a detailed benchmark that also included gcc (I think it was 4.5?)... but it was so far off the pace, I havent really considered it a possible contender since.

now, to be fair, there is a huge "your mileage may vary" effect here.

The code I use for benchmarking is almost pure C and has lots of threading, lots of SSE intrinsics, no virtual function, no STL, no RTTI, no exception handling, and I've done lots of hand optimizations to help coax the compilers to make better output.

yeah, there are a number of different optimization settings between the different compilers... so you really should take at least a little bit of time with the compiler settings in order to make a fair comparison.

So, if you tell me that you found gcc to outperform intel 13, then yeah, I'll raise an eyebrow and say "o rly?"... but then if you say "ya rly", I'd probably believe you.


RE: c++ hate...

C++ is just so old and has grown so big, there's simply something for everybody to hate about it. It certainly gives the programmer more rope with which to get tangled up in and/or hang themselves than any other language that I can think of.

That said, despite all the things to complain about, for many kinds of projects and even some entire industries, there is simply no alternative AINEC.

I agree pretty much with Neil S. One thing I think is pretty clear, certain languages lend themselves to certain problem domains.

Still too much reading. Just look at the pretty pictures: http://yosefk.com/c++fqa/linking.html

irt C++. Not interested in learning it at this moment. I took a little look at some code. I would think I'd be able to recognize something but I pretty much got stuck at the header includes. Totally different language.

The way you guys describe it, C++ would be like someone picking up "Javascript, The Good Parts," reading the entire "bad parts" section, dismissing the "good parts," and creating Javascript++.

Eh, one day there will be that "perfect" language.

I felt the same way about doing courseware at first. You are given programs that are generally 15 LOC and then you move on to the next toy problem. These toys feel small and it feels like you really aren't learning anything "useful."

At some point, you'll want to put your knowledge to the test and the real measure will be a simple question: "Had I attempted this one year ago, would I have been able to solve this problem with the elegance that I have done so today?" When you think about it, you'd probably realize that you probably would not be able to, and all those stupid toy programs really end up being extremely valuable. A larger program is partly a result of tacking small solutions onto small solutions, which you presumably became good at after doing a thousand small problems.

You may not see the value of figuring out why you need to compute Newton's Approximation in the real world, but trust me, there is a valuable lesson hidden inside of there that probably won't hit you until you need the /lesson/ it was showing you.

dave,

curious why you think C is "totally awesome"? especially as someone who has experience with clojure and python....

fwiw, i didn't get that at all (but i do see why my response could be read as implying that, even though it wasn't my intention)

Anyone have experience with programming UCMA applications?