umd

I don't remember there being a huge amount of difference between the editions.

kudude

fortran anyone?

tallmantim

And if you put a really big page number at the end it will just take you straight to the last page:

https://www.bikeforums.net/showthread...24026/page9999

KiddSisko

Sub wrote elsewhere:
You linked to a blog post written in '09. There hasn't been progress in dealing with it since then? And by dealing I mean a simple and effective way of removing it from a system.

mollusk

If you are really old it is FORTRAN.

I used to program Fortran IV a lot. On cards. And I had my own "drum" for the keypunch machine. I wonder where it is now.

The CFD guys still like the modern versions.

kudude

fortran95 is fortran. FORTRAN77 is FORTRAN. Thank god I don't have to deal with all the asinine rules regarding whitespace. Don't know what CFD is, but when it comes to matrix multiplication and large parallel jobs, fortran makes me

umd

Assume it's Computational Fluid Dynamics.

The core of our software was all written in fortran and converted to C eons ago.

agoodale

In my first Jr. high programming class we had to bubble cards in FORTRAN. We had a whole lab of new Apple IIs. Our teacher didn't have any curriculum for them so we used them for fantasy baseball instead.

wens

Fun stuff. I had a professor who wrote a program in CFD to model missiles for star wars. We convinced him to tell us a little about it while we were driving to Wright-Pat, it was really interesting.

Urthwhyte

and my latest project was only figuring out how to optimize caching infrastructure. I need to grab a book on R or F# and start working on something more interesting than Web App #1238021.

kensuf

you know guys, there's a twitter feed to post the talk about compilers and K&R etc..

Urthwhyte

Are you kidding me? Discussing K&R takes far more than 140 characters. Discussing pointers alone takes at least .8cdr

kensuf

pointers are the reason I quit programming c and never turned in my final project for cis 2321.

waterrockets 

Well, it could be the other way around where I can't get Dalvik to GC some big allocations. I haven't dug in too deep yet, but I'd much rather deal with pointers than try to get a Java VM to give me back my memory when I'm running out on a smartphone.

Back when I was working on games, I did a stint as "the memory guy," chasing allocations in C++. Going after inefficiencies, leaks, and my favorite: memory fragmentation. Putting the beatdown on some fragmentation is actually really fun.

slynkie

+1...

though it seems whenever i'm working with Java, I miss C++. whenever I'm working with C or C++, I reluctantly miss Java.

I never miss assembly.

slynkie

chasing leaks + inefficiencies are somewhat normal to me now, and there are some good tools out there. i'm not sure i'd even know where to begin to hunt down fragmentation though. got any good primers/reading recommendations?

jwible

Well at least Java doesn't have pointers as evidenced by the utter lack of NPEs when testing code.

waterrockets 

I never read anything about it, but we had a great logging tool. We'd written our own memory manager and all allocations could be logged with a time stamp. In our heap view, we could "turn back time" to see what the heap looked like a different points in the log. So, I'd look at the heap, find medium-sized chunks of free space, back up until allocations filled it, and check the allocation. Then we could move those allocations around, delaying their loading or whatever had to be done, so that when the memory was released, it hadn't pushed out other significant allocations.

Without that time-travel logging tool, it's just a matter of using a heap analysis tool and looking at it after the system has been running a while and is stable. You look for gaps beneath larger allocations, then do a binary-style search by restarting and checking the heap at different points in the startup to see where that block was still allocated, then check the call stack.

Another trick is to write a message to your memory before you release it for any large allocations being performed. So, for our game intro video, for example, we could write 0xDEADBEEF in every slot before calling delete/free(). Then, in heap analysis, if we saw a fragmentation gap that was much smaller than the intro video, but still a frag issue, we could identify it as the video by seeing 0xDEADBEEF in there. You can choose different messages for different software components or major allocations.

Once you start beating on this stuff, little tricks like this will pop into your head. You develop a Sherlock Holmes attitude, and a bit of Spidey Sense about it. "My spidey sense is tingling -- I think it's MusicManager.init()!!"

waterrockets 

Yeah, I'm pretty language agnostic. I like them all. I really like learning new languages, but you start to run out... of course I never remember them due to the "use it or lose it" clause.

ridethecliche

Hai gaiz.

4chan anyone?

slynkie

thanks for the description - that does sound fun. that type of logical hunt can be very satisfying.

substructure

extruded plastic dingus

kensuf

that's what she said.

Spring Water

I've got a pretty good handle on PHP, at least enough to build some pretty rudimentary scripts and enough to edit someone elses to make it to what i need it to. That's all well and fine for web-apps, but I've got some ideas that would require a more powerful language.

Which direction would you point the self-learner to start?

Scummer

Scary, when I actually understood what they were talking about the "virus" that supposedly infected the Iranian uranium enrichment. I've programmed in SIMATIC Step5 for a few years before they switched to a C like Step 7. In my case controls for hydro power plants, so I had nothing to do with the Iranians.