|
Thread Rules
1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution.
2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20)
3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible.
4. Use [code] tags to format code blocks. |
I'm not sure -- I just don't think that you can look at a class and treat it as a method list and nothing else. To start with, those methods all have names; surely you agree that the names are important? Suppose Array.BinarySearch always returned the right result, but it only did it in linear time; wouldn't you be surprised and irritated that it wasn't logarithmic?
How do you know whether a piece of data ought to be exposed as a property or a method call in C#? Usually, the convention is that if it could take a non-negligible amount of time to get the result, it ought to be a method call (barring some lazy initialization sort of hijinx.) So here is already an example where C# suggests different best practices based on the performance of your implementation.
Most C# programmers who are collaborating with other people are vaguely familiar with most of the framework classes that implement IList and ISet, so it's not a matter of looking at all the concrete classes; you've already seen them. I think that real efficiency is being able to quickly gauge a big piece of code based on the shared experience and conventions between you and the writer, maximizing the amount of information you're communicating with your choices of design, variable names, and types. I don't see any reason that performance characteristics shouldn't be part of that information.
|
On July 13 2010 00:51 Adeny wrote:See how confusing that is?  Think I got it though, checking out LINQ and it's some really funky stuff. var lowNums =
from n in numbers
where n < 5
select n;
SQL-syntax for C#...
LINQ is a pretty damn nifty feature. There's a nice blog article pertaining to the cartesian product you're essentially after, but I'll have to get you the link tomorrow at work. It's a bit more than what you need, but it was an interesting read. The jist of it, iirc, was how to go about performing the cartesian linq when you don't know how many sets you have.
|
Edit: seeing what is there and what is actually there is probably the worst part of debugging, to be honest.
|
Presumably because he doesn't want to return false and instead keep looping.
|
On July 13 2010 15:12 Tyraz wrote:
Edit: seeing what is there and what is actually there is probably the worst part of debugging, to be honest.
What do you mean? Are you debugging with optimizations on or not using a modern debugger? 
|
I'm looking to get into parsing binary information such as SC replays and maps. I've read through the source of some parsing projects and to me it seems as if they are looking for specific codes (memory codes?) in the file which map to an event/action in plain english. Could someone explain this to me?
I also don't know what words to search for when trying to find more information on this subject. I have tried binary file parsing to no avail. Any ideas or links to docs/tutorials?
Thanks in advance for any help.
|
Just a small note: The link to w3schools in the OP is wrong. I guess it should be .com instead of .org.
|
Adeny:
You piqued my interest sufficiently in Project Euler that I finally went for it. I did project #23 and enjoyed both the process and the afterstudy. Looking at other's code for the same problem is a telling experience -- sometimes things look so similar it's absurd, and likewise, sometimes so different it's equally absurd. Throw in people with different language preferences and the variety can be astonishing. I had a reasonably fast method but no matter how clever you think you are there's always someone who optimized it in ways you never considered. Mine ran in maybe half a second after I finished optimizing.
Key points for speed: 1) improve your abundant-number check by iterating only to square root of n, and whenever you find a match, count it for both values of the divisor-pair. 2) Use a system other than foreach to loop through the array, since that will compare every value in the array to every other value twice. This could also improve your early-exit condition by testing for if( i + ii > counter ) instead of each one individually.
All: pretty awesome that this thread has come about. Coders from all types frequent here, seeming to almost come out of the woodwork at times. I've browsed TL for a long time and have always noted that there are some serious developers around. More and more people are getting into it though which is rather nice. When I manage to get my thoughts in order, I plan to start up some discussion regarding three areas: DLL injection and other stealth techniques, dynamic web scraping, and OCR/screenscraping. I'll be able to offer a degree of insight into such things and then hope to take full advantage of those who know more than I do, perhaps ultimately being able to glean some advice or get some questions answered.
|
|
|
On July 13 2010 18:59 lastkarrde wrote:
I'm looking to get into parsing binary information such as SC replays and maps. I've read through the source of some parsing projects and to me it seems as if they are looking for specific codes (memory codes?) in the file which map to an event/action in plain english. Could someone explain this to me?
I also don't know what words to search for when trying to find more information on this subject. I have tried binary file parsing to no avail. Any ideas or links to docs/tutorials?
Thanks in advance for any help.
lastkarrde:
Ahh, data file parsing. The beauty, as a developer, of making your own data file formats for your programs is that you can create them however you see fit. This same beauty can prove very, very ugly for those of us trying to interpret it without the documentation of what anything means. For example, let's look at a Move command in a Starcraft replay. The important parts of the Move command are the command title, MOVE, the unit(s) to move (probably represented by an ID number), and a destination (in x-y coordinates, perhaps normalized such that 0,0 is one corner and 65536, 65536 is the other). In English, then, this command could be represented by:
Move Marine (UnitID:62) to location 2640, 13280.
This however, is a lot of text. What if instead of writing Move, we just have a symbol that means move. Say, the characters MV. And instead of listing the extra stuff for the unit, we just list the ID (since there is only one Unit#62). And let's take out the words 'to location' as well. The computer doesn't need that, it's assumed from a move command. We are left with:
MV 62 2640 13280
Now, that's close to how data is actually stored. However, instead of using nice things like MV for move, or actual numbers, pretty much everything is converted to Hex. Actions are probably represented something like this
+ Show Spoiler +
00 - Stop
01 - Attack
02 - Psi storm
03 - Hallucinate
...
09 - Switch to Unseige mode
0A - Switch to Seige mode
0B - Lay spider mine
..
0F - Parasite
10 - Ensnare
11 - Disruption Web
...
9F - Gather
A0 - repair
A1 - lift off
A2 - archon meld
A3 - move
A4 - build interceptor
...
etc
Perhaps that's overkill, but it helps give you the idea. It might even be possible that there's so many commands (more than 256), that they are all represented by a 4-digit hex string. Anyway, in this example, move is A3. The Unit ID (#) in hex would be 3E. The location x,y would be 0A50, 33E0. Now, also, everything is padded with zeroes to be fixed length, so no punctuation is needed. So here's our new funky command:
A33E0A5033E0
There's a lot of other stuff that could be going on, such as variable length commands (if there are multiple units moving somewhere, for example), 'punctuation' codes (maybe all commands end with 0F0F or something). This should give you a taste though.
To write a program that reads these files, you need to basically reverse-figure-out how the developer decided to write everything. It can be anything from a minor hassle to an enormous undertaking. I'd imagine with replay/map files it's somewhere in the middle -- but a bit closer to an enormous undertaking than a minor hassle.
|
Spoiler: Project Euler 89 solution
+ Show Spoiler +http://pastie.org/1042721
To whoever was interested in text parsing, take a look at this. It's a very basic text parsing program. It's also somewhat efficient, runs in 15ms. I see quite a bit of room for optimization but oh wells.
Edit: Forgot to add it's in C#, which I have only been messing with for about a day now so don't expect much.
|
On July 14 2010 02:18 Adeny wrote:Spoiler: Project Euler 89 solution + Show Spoiler +http://pastie.org/1042721 To whoever was interested in text parsing, take a look at this. It's a very basic text parsing program. It's also somewhat efficient, runs in 15ms. I see quite a bit of room for optimization but oh wells. Edit: Forgot to add it's in C#, which I have only been messing with for about a day now so don't expect much.
A couple quick tips to let you know about C#-specific stuff that you might not realize:
* You can index into a string to get an individual character, just as if it were already an array of characters. So you don't need to copy it into a buffer to look at it.
* If you're using .NET 4.0, your file reading code could look more pleasant, like this:
foreach (string line in File.ReadLines(path))
sum += /* ... */
If you're on 3.5 or earlier, you can use File.ReadAllLines to do the exact same thing, but ReadAllLines reads the whole file into memory at once so it may not be so hot for a big file.
|
I'm about to redesign an entire system for the company i work for (a small warehouse company) and am interested to know if anyone uses .NET MVC 2 and how it stacks up against other web frameworks like ruby on rails or Zend.
|
On July 14 2010 02:41 catamorphist wrote:Show nested quote +On July 14 2010 02:18 Adeny wrote:Spoiler: Project Euler 89 solution + Show Spoiler +http://pastie.org/1042721 To whoever was interested in text parsing, take a look at this. It's a very basic text parsing program. It's also somewhat efficient, runs in 15ms. I see quite a bit of room for optimization but oh wells. Edit: Forgot to add it's in C#, which I have only been messing with for about a day now so don't expect much. A couple quick tips to let you know about C#-specific stuff that you might not realize: * You can index into a string to get an individual character, just as if it were already an array of characters. So you don't need to copy it into a buffer to look at it. * If you're using .NET 4.0, your file reading code could look more pleasant, like this: foreach (string line in File.ReadLines(path))
sum += /* ... */ If you're on 3.5 or earlier, you can use File.ReadAllLines to do the exact same thing, but ReadAllLines reads the whole file into memory at once so it may not be so hot for a big file.
How do I go about indexing into a string? I know strings are essentially char arrays, but szHi[5] doesn't work at all. I tried foreaching through a string, but I don't think this would work;
foreach (char c in szMyString)
{
if (c+1 == 'M')
}
Because from what I can gather char c wouldn't be an index, but rather the value of the char. The foreach for reading the file looks a bit better though, don't know if it's different performance wise.
|
I don't know quite what you're observing, this should work:
string s = "fatso"; char ch = s[2]; // ch is 't'
See http://msdn.microsoft.com/en-us/library/362314fe(VS.71).aspx for details.
You're right that foreach iterates over values and doesn't let you get at the index, so that wouldn't work out. You would be best off with a for loop just as you have it.
File.ReadLines uses a StreamReader internally in a loop, just like you've written it, so there shouldn't be any noticable performance differential at all.
|
Sorry you're right, I don't know how I got the string stuff mixed, maybe I thinking C/++ for a minute there.
|
On July 14 2010 03:51 Adeny wrote:
Sorry you're right, I don't know how I got the string stuff mixed, maybe I thinking C/++ for a minute there.
C++ strings allow access via operator[] too, and foreach in C++ is a template that basically takes a container class and a functor. You should not confuse for(int i = first ; i != last ; ++i) with a foreach loop. You can only use it to emulate one, but it is something different.
|
On July 13 2010 18:59 lastkarrde wrote:
I'm looking to get into parsing binary information such as SC replays and maps. I've read through the source of some parsing projects and to me it seems as if they are looking for specific codes (memory codes?) in the file which map to an event/action in plain english. Could someone explain this to me?
I also don't know what words to search for when trying to find more information on this subject. I have tried binary file parsing to no avail. Any ideas or links to docs/tutorials?
Thanks in advance for any help.
RoTaNiMoD already posted a great explanation of the way to approach data file formats, but I was curious if you meant SC as in original or maybe SC2. I wrote the sc2 map analyzer which has a lot of data file parsing to do, and if that's where you at then you should definitely check out all the information the dudes at SC2Mapster.com figured out about the internal map formats. They have a section for replays but it looks like nobody has started pulling those apart over there yet. In that case you might want to PM Dakota_Fanning here at TL who wrote the sc2gears tool to rip open replays.
It's fun stuff, and I hope you get your project going.
|
On July 14 2010 04:02 spinesheath wrote:Show nested quote +On July 14 2010 03:51 Adeny wrote:
Sorry you're right, I don't know how I got the string stuff mixed, maybe I thinking C/++ for a minute there. C++ strings allow access via operator[] too.
*Mind blown*. I have been wrong all my life... How did I miss this?
THE HOOPS I HAVE HAD TO JUMP. THE ARRAYS I HAVE HAD TO CAST. Etc.
|
I think it would be cool if we had a poll on the first post in the thread that allowed you to vote for your favourite language. I would be really interested in that.
|
|
|
|
|
|
EPT
