Nasty code...but please critique it anyway :-)

[Michael Strorm]

Hi!
I posted a message a while back asking for project suggestions, and decided
to go with the idea of creating an adventure game (although it was never
intended to be a 'proper' game, rather an excuse to write- and learn- some
C++).

To cut a long story short, I wrote a fair chunk of it, but realised that
it's... not very good. Okay, it's my first "proper" C++ program, so that's
no big deal, but I don't want to waste more time working on a program that
should be rewritten from scratch (I want to start reading 'Accelerated C++'
instead). What *would* be interesting would be to hear what other people
think about the (compilable, but not properly working) code.

It's a long program, spread over many files, so I haven't posted them here.
The URL is
http://www.mstrorm.free-online.co.uk/
Yes, I know the code isn't particularly well laid out or commented- but I
hope it's clear enough.

The design of the program is that a Controller class controls the flow of
events and oversees everything. A base-class 'Noun' includes common behavior
and is extended to give us game locations, items (i.e. physical objects),
Beings (i.e. game characters, further subclassed for the Player object which
represents the human player) and exits.
Given a command such as
eat chocolate
or
go north
the controller finds the Noun-subclass-object representing chocolate (an
Item) or north (an Exit)- that object then does the action "eat" or "north"
respectively. Assuming the Noun-subclass supports that action, it can
respond appropriately (e.g. the Exit object with id=="north" would update
the player's location in its "go" method to wherever the exit pointed to).
Note that action methods (or, more specifically, functions) are static
functions held in a vector of pointers. I don't like the way that the vector
of function-pointers is initialized, but it seemed the best way at the time.

Criticisms of the large-scale design, or basic programming would be equally
useful. I think there's plenty to criticize about both, as you can see from
the comments I left in (though I like the subclassing of Nouns into
different object types).

Anyway, feedback would be appreciated....

 

[Martijn Lievaart]

It's a long program, spread over many files, so I haven't posted them here.
The URL is
http://www.mstrorm.free-online.co.uk/
Yes, I know the code isn't particularly well laid out or commented- but I
hope it's clear enough.

Better create a zip and/or tarball, that's much easier to download.

HTH,
M4

 

[Jerry Coffin]

[ ... ]

The design of the program is that a Controller class controls the flow of
events and oversees everything.

IMO, you've put too much into your Controller class -- a search on
Google (in the newsgroups) for "God class" should turn up some
interesting reading on how to de-centralize the control a bit.

A base-class 'Noun' includes common behavior
and is extended to give us game locations, items (i.e. physical objects),
Beings (i.e. game characters, further subclassed for the Player object which
represents the human player) and exits.

IMO, "Noun" is so general as to be of little use. Looking through
Noun.hpp, it has (for example) a getLocation and setLocation -- but
looking at Locn.hpp, a Locn _is_ a Noun. This doesn't make a lot of
sense to me.

Given a command such as
eat chocolate
or
go north
the controller finds the Noun-subclass-object representing chocolate (an
Item) or north (an Exit)- that object then does the action "eat" or "north"
respectively. Assuming the Noun-subclass supports that action, it can
respond appropriately (e.g. the Exit object with id=="north" would update
the player's location in its "go" method to wherever the exit pointed to).

I don't think this is really a very good way of structuring things. A
person is giving a command to their player, so the player object should
receive the command. The player object should probably have a parser
object to figure out what the command means. Based on that, it should
check its environment and try to figure out a way to carry out the
command.

I don't think it makes much sense for a location to derive from the same
base as an item -- and unless you want (for example) to allow for the
possibility of a player carrying around another player like it would
carry around a sword, goblet, etc., it probably doesn't make much sense
for a player and a goblet to derive from the same base either.

If I were doing this, at the top level I'd have some locations and some
beings. Each location contains a collection of exits, another of items,
and possibly a set of attributes (e.g. a particular room might be dark
so the player can't see in it). Each location will also have a
collection of beings that are within that room at the present time.

Each being has a current location, a collection of items it's carrying,
probably some other attributes to describe its strength, talents,
skills, etc.

A player is a being that also have a command parser so it can take
orders. The commands go directly to the player, which then uses its
command parser to figure out what it is. Based on its current
environment (skills, strength, environment, etc.) it tries to figure out
a way to carry out the command its received. e.g. if you say "go
north" it looks at the collection of exits in its current location, and
sees whether there's an exit to the north. If so, moves to that
location (updating its current location and updating the locations'
collections of what they contain).

I also wouldn't hard-code the commands to which a player can respond.
Instead, the player data file contains data telling about what commands
a player can respond to, and what the effects of that command are.
Likewise, each item would have a set of commands that it responds to (or
enables) and the effects of those as well. Just for example, you might
decide that a player can fly if he has a talent for a particular type of
magic OR if he is wearing a particular cloak. This would enable a
player to respond to a command to go up that would otherwise not be
allowed.

This would make a relatively easy to add (for example) new food items to
your game, each with different effects on strength, reaction speed, and
so on. Likewise with adding new weapons, monsters, etc.

 

[Michael Strorm]

Better create a zip and/or tarball, that's much easier to download.

Sorry- should have mentioned that there's a zip of the whole lot somewhere
in there!

 

[Martijn Lievaart]

Sorry- should have mentioned that there's a zip of the whole lot somewhere
in there!

Looked at the listing and missed it. Damn! Sorry about that.

M4

 

[J. Campbell]

I, too, would like feedback on my code.
I'd appreciate feedback on:

1) my use header files
2) the inclusion of namespace std
3) the general organization of the program
4) anything else that you might care to comment on

The source is located in J_Crypt.zip at:
http://home.bellsouth.net/p/PWP-brightwave

J_Crypt a file encrypter/unencrypter that is based on a PRNG I
designed (contained in slicerclass.h and slicerclass.cpp). This
project was an excuse to get started with c++.

To use the program, you "drop" a file onto the executable and follow
the menu to encrypt or unencrypt the file. The program is set up to
work with dos-style paths:
drive:\path\filename.extension

if another system is used, filespec_class.h and filespec_class.cpp
will need to be modified to appropriately parse the commandline.

Thanks for taking the time to look at this.

 

[Michael Strorm]

[ ... ]

The design of the program is that a Controller class controls the flow of
events and oversees everything.

IMO, you've put too much into your Controller class -- a search on
Google (in the newsgroups) for "God class" should turn up some
interesting reading on how to de-centralize the control a bit.

I'll take a better look for that later on; it sounds interesting (and pretty
useful), but I wanted to reply to this post first. It definitely helped when
I blocked the obvious religious words BTW.

IMO, "Noun" is so general as to be of little use. Looking through
Noun.hpp, it has (for example) a getLocation and setLocation -- but
looking at Locn.hpp, a Locn _is_ a Noun. This doesn't make a lot of
sense to me.

I agree; I'd thought about that myself. However, a location, like an Exit,
Item or Being, can be the 'noun' in the typical 'verb noun' or 'verb noun
preposition noun2' input forms (eg "look room"), and to make it a non-noun
would mean rewriting/rehashing of existing code with additional handling
required, plus twice as much maintenance. That's why I went with the noun
solution even though, as you say, there are some inconsistencies, such as a
location having a location. The other thing I wanted to avoid was runaway
child classes for every minor variation (which would have solved the
inelegance above at the expense of more classes). I'm not disagreeing with
your criticism, only trying to explain why it was written that way.

to).

I don't think this is really a very good way of structuring things. A
person is giving a command to their player, so the player object should
receive the command. The player object should probably have a parser
object to figure out what the command means. Based on that, it should
check its environment and try to figure out a way to carry out the
command.

My reasoning here is that the player doesn't need to know about everything
it can carry out an action on- of course, dependency is hard to avoid (if
player eats an apple, the apple might have to know about the player if
player is a special case of being.... this could go on).

I don't think it makes much sense for a location to derive from the same
base as an item -- and unless you want (for example) to allow for the
possibility of a player carrying around another player like it would
carry around a sword, goblet, etc., it probably doesn't make much sense
for a player and a goblet to derive from the same base either.

Perhaps I'm thinking too hard about simplifying the parser. It just seemed
very OO to let an object determine its own behaviour rather than what was
carrying out the action- making it easier to expand.

If I were doing this, at the top level I'd have some locations and some
beings. Each location contains a collection of exits, another of items,
and possibly a set of attributes (e.g. a particular room might be dark
so the player can't see in it). Each location will also have a
collection of beings that are within that room at the present time.

Yeah, this is where I can see the point you made about the god-class above.
I'm not sure that the controller needs to be able to handle that kind of
thing- it wasn't part of the design I put that much thought into, on
reflection.
Question... would pointers in both directions (Player having *Locn and Locn
having *Player) be a good idea? This is what you seem to suggest below...

Each being has a current location, a collection of items it's carrying,
probably some other attributes to describe its strength, talents,
skills, etc.

I also wouldn't hard-code the commands to which a player can respond.
Instead, the player data file contains data telling about what commands
a player can respond to, and what the effects of that command are.
Likewise, each item would have a set of commands that it responds to (or
enables) and the effects of those as well. Just for example, you might
decide that a player can fly if he has a talent for a particular type of
magic OR if he is wearing a particular cloak. This would enable a
player to respond to a command to go up that would otherwise not be
allowed.

Ideally, I'd like to have very little hard-coded. The way I wrote it was
intended to be a compromise between code-flexibility and hard-coding
(hard-coding making it easier to customise behaviour, but not being as
elegant or convenient).

This would make a relatively easy to add (for example) new food items to
your game, each with different effects on strength, reaction speed, and
so on. Likewise with adding new weapons, monsters, etc.

Would it be possible to add custom code for each item, or just vary a few
(pre-decided) parameters?
IIRC Java can load classes dynamically, but even if ISO-C++ can do this, I
don't think my knowledge is good enough to do that yet.

With the existing code, what I don't like about my customised Noun behaviour
in particular is (a) The C-style nature of the code, and (b) Use of static
functions (since, apparently, a static function within the class still has a
different signature depending on the class, that is, function of type foobar
has type Noun::foobar if defined within Noun, and Being::foobar if defined
within Being, leading to PITA compatibility problems and my kludgey
solution).

Anyway, interesting answers, thanks!

- MS

 

[Jerry Coffin]

I, too, would like feedback on my code.
I'd appreciate feedback on:

1) my use header files

Not really very good -- each of your .cpp files should include the
header declaring the class(es) for which it contains implementation.
As-is, I'm not at all sure how you got some of the .cpp files to compile
at all.

2) the inclusion of namespace std

Well, you certainly don't make what I'd consider optimal use of the
standard anyway. Consider (for one example) these three functions:

string lcase(string in){
string stringout;
for(int i = 0; i < in.size(); ++i)
if(!(in & 128) & ((in & 95) > 64) & ((in & 31) <= 26))
stringout += (in | 32);
else stringout += in; //character wasn't a letter...dont change
return stringout;
}

string ucase(string in){
string stringout;
for(int i = 0; i < in.size(); ++i)
if(!(in & 128) & ((in & 95) > 64) & ((in & 31) <= 26))
stringout += (in & (223));
else stringout += in; //character wasn't a letter...dont change
return stringout;
}

These are barely understandable at all. I guess they're intended to
convert entire strings so all letters are upper case or all characters
are lower case. For that sort of job, I'd do something like this:

typedef std::string::size_type s_t;

std::string lcase(std::string in) {
std::string ret;

s_t len= in.size();

ret.resize(len);

for (s_t i=0; i<len; ++i)
ret = std::tolower(in);
return ret;
}

and similarly for ucase. Better yet, you could use std::transform
instead of an explicit loop:

std::string lcase(std::string in) {
std::string ret;

std::transform(in.begin(), in.end(),
std::back_inserter<std::string>(ret), std::tolower);
return ret;
}

std::string ucase(std::string in) {
std::string ret;

std::transform(in.begin(), in.end(),
std::back_inserter<std::string>(ret), std::toupper);
return ret;
}

Likewise, with this:

string center(string s){
int p = s.size();
string temp;
if(s.size() < 80){
for(int i = 0; i < (39 - (p/2)); ++i)
temp += ' ';
}
return (temp + s + "\n");
}

First of all, this assumes you're only ever going to center things over
an 80 column area, which is pretty inflexible. Second, it does that
with less than, shall we say, the greatest of aplomb. I think I'd use
something like this:

std::string center(std::string s, int width = 80) {
std::string ret;
std::fill_n(std::back_inserter<std::string>(ret),
(width-s.size())/2,
' ');
return ret+s+"\n";
}

Or perhaps:

std::string center(std::string s, int width = 80) {
std:stringstream os;

os << std::setw(width - (s.size()/2)) << s << "\n";
return os.str();
}

3) the general organization of the program

It looks to me like you need to work on generalizing things. Just for
example, your "gun" does't just read input -- it also reports errors (to
standard output) so it can only ever be used interactively.

It's usually better to have one part that does only reading, and if it
fails, reports the failure only in a return code, by throwing an
exception, etc. (I.e. only to other parts of the program, NOT directly
to the user). Then, a separate part handles the direct interaction with
the user, based on the feedback from the internal function.

This generally improves portability (and often readability), and makes
each easier to transplant into other situations with a total rewrite.

As far as overall organization, it may simply be that I'm tired right
now, but I couldn't follow it at all. As I said above, I'm not even
quite sure how some of it compiles...

 

[Chris \( Val \)]

| In article <[email protected]>,
| (e-mail address removed) says...

[snip]

Hi Jerry.

| typedef std::string::size_type s_t;
|
| std::string lcase(std::string in) {
| std::string ret;
|
| s_t len= in.size();
|
| ret.resize(len);
|
| for (s_t i=0; i<len; ++i)
| ret = std::tolower(in);
| return ret;
| }

[snip]

It is very rare that I see anyone using locales
with 'std::tolower()', etc. For example, the
above might have been implemented as follows:

# include <locale>
std::tolower( in, std::locale() );

....but the real question(s) are:

Why do many people choose not to use the locale
version ? - an loaded question, I know .

Is there no advantage to the locale version ?

I was always under the impression that the locale
version offered better safety, due to the possible
use of different character sets being manipulated.

Comments ?

Thanks.
Chris Val

 

[J. Campbell]

Not really very good -- each of your .cpp files should include the
header declaring the class(es) for which it contains implementation.
As-is, I'm not at all sure how you got some of the .cpp files to compile
at all.

Thanks...I knew I wasn't quite doing it right...So...I should be
including *.cpp files that have the *.h declarations included at the
top rather than including the *.h files with the *.cpp definitions
included at the bottom. This is clear.

Well, you certainly don't make what I'd consider optimal use of the
standard anyway. Consider (for one example) these three functions:

It's ironic...you focused on "non-important" helper functions that I
wrote to get input or for text formatting. I agree that your methods
are better. I wrote the case-functions some time back while looking
at the ascii table layout. It was an excercise to see what to_upper
and to_lower really involved...anyway...I'm not making excuses.
Thanks for taking the time, and getting me straightened out on the
header-file stuff.

It looks to me like you need to work on generalizing things.

Thanks. I agree.

As far as overall organization, it may simply be that I'm tired right
now, but I couldn't follow it at all.

Don't bame your tiredness...I'll accept the responsibility...and work
on improving readability.

As I said above, I'm not even
quite sure how some of it compiles...

It acctually compiled fine with dev-c++. I think that my reversing
the *.h and *.cpp included causes confusion. Still, I have
function/class declarations, followed by their
definitions/implementations...

Anyway...Thanks a bunch for you time and help. I have much to
learn...and it's really useful to get human feedback rather than the
only feedback coming from the compiler!!

 

[Jerry Coffin]

[ ... ]

Why do many people choose not to use the locale
version ? - an loaded question, I know .

I can't speak for anybody else, but in my case, because it's extra work
(albeit not a great deal).

Is there no advantage to the locale version ?

There is an advantage under some circumstances, but most people don't
encounter those circumstances very often. In code like I was posting,
it would only have been an unnecessary distraction in any case.

I was always under the impression that the locale
version offered better safety, due to the possible
use of different character sets being manipulated.

Less a matter of safety than versatility. If you don't pass a locale,
then they default to a single global locale. When/if you have to deal
simultaneously with input in a number of different character sets, then
an explicit locale becomes very handy. Until or unless you encounter
that, it's of little real consequence.

 

[Jerry Coffin]

[ ... ]

It's ironic...you focused on "non-important" helper functions that I
wrote to get input or for text formatting.

This was largely because in those cases the names of the functions made
it fairly clear what they were intended to do, so it didn't take much to
rewrite them. I leave it to you to understand your own encryption
algorithm sufficiently to apply the same principles to it.

[ ... ]

Anyway...Thanks a bunch for you time and help. I have much to
learn...and it's really useful to get human feedback rather than the
only feedback coming from the compiler!!

'tis true: compilers are pretty short on advice about style
(intentionally so, to a large extent).

 

[Agent Mulder]

I, too, would like feedback on my code.
http://home.bellsouth.net/p/PWP-brightwave

</>

Hey Joe,

I like your code. I only gave it a brief look, since I am
busy programming on something else, but I'll give you
my comments:

-You have a funny style. Files with only one function in
it. That is ok. Try everything.

-Stick to having one file in every project (or directory) called
main.cpp. It contains the main routine, as expected, but also has
a special importance in program organisation. Your
'main.cpp' is called J_Cript.cpp. That is too cryptic a name.

-Merge the .h files and the .cpp files. Put one class declaration
+ the method definitions in one file. You can inline everything if you
like (I do it only), but that is un-C++-ish. You can also define
the methods just below the class declarations, still in the same file.

-Step 3. You now have a 1-to-1 relation between a file and
a class. The extension of this file is arbitrary. It is both a .h
file with a declaration and a .cpp file containing the definitions.
Extend it with h.

-In file main.cpp, write out the #included .h files like this:

#include <iostream>
#include <iomanip>
#include <fstream>
#include <string>

using namespace std;

const string version = "J-Crypt 01.0.0.2";

#include "AlignInt_class.h"
#include "describe.h"
#include "Encrypt.h"
#include "Filespec_class.h"
#include "hashSclass.h"
#include "JKC_functions.h"
#include "J_Crypt.h"
#include "slicerclass.h"
#include "unencrypt.h"

int main(int argc, char* argv[]){
cout << "\n" << center(version) << center("2003 Soft Corp.") << endl;
if(argc < 2){describe(); wait(); return 0;}
....

-You now have all your class enumarated in file main.cpp. You can
remove all and every #include in any other file. Really. This has some
major advantages (and a drawback).

Advantages:
- You can remove all and every #include in any other file
- You have a list of all the classes used by your program ready in main.cpp
- You are forced to have the hierarchy of your objects right, thus reinforcing
the design.

Drawback:
- Everything gets compiled all the time.


That is how I enforce rigidity in my programs. Try it your way.

-X

 

[Chris \( Val \)]

| In article <[email protected]>, "Chris \( Val
| \)" <[email protected]> says...

Hi Jerry.

| > Why do many people choose not to use the locale
| > version ? - an loaded question, I know .
|
| I can't speak for anybody else, but in my case, because it's extra work
| (albeit not a great deal).
|
| > Is there no advantage to the locale version ?
|
| There is an advantage under some circumstances, but most people don't
| encounter those circumstances very often. In code like I was posting,
| it would only have been an unnecessary distraction in any case.
|
| > I was always under the impression that the locale
| > version offered better safety, due to the possible
| > use of different character sets being manipulated.
|
| Less a matter of safety than versatility. If you don't pass a locale,
| then they default to a single global locale. When/if you have to deal
| simultaneously with input in a number of different character sets, then
| an explicit locale becomes very handy. Until or unless you encounter
| that, it's of little real consequence.

Thank you for clarifying that, it makes sense .

Cheers.
Chris Val

 

[J. Campbell]

Not really very good -- each of your .cpp files should include the
header declaring the class(es) for which it contains implementation.
As-is, I'm not at all sure how you got some of the .cpp files to compile
at all.

I must be pretty dense...I had the hardest time understanding what I
was doing wrong. I think I get it now. Will you take a look at the
(revised) code from my first link in this thread and tell me if my use
of headers is now more or less normal?
int main(){} is located in j_crypt.cpp

thanks again for the response.

 

[Jerry Coffin]

[ ... ]

I must be pretty dense...I had the hardest time understanding what I
was doing wrong. I think I get it now. Will you take a look at the
(revised) code from my first link in this thread and tell me if my use
of headers is now more or less normal?

If I still had the link handy I would, but I don't...

 

[J. Campbell]

[ ... ]

I must be pretty dense...I had the hardest time understanding what I
was doing wrong. I think I get it now. Will you take a look at the
(revised) code from my first link in this thread and tell me if my use
of headers is now more or less normal?

If I still had the link handy I would, but I don't...

http://home.bellsouth.net/p/PWP-brightwave

Sorry I didn't include the link before...I didn't want people to think
that I was entheusiastically pimping my novice code, and the link was
just up the thread, at least on my newsreader ;-) Anyway, thanks for
responding.

 

[Jerry Coffin]

Well, I've taken a look at the code, and I still think it could use some
work. One obvious point would be that there are quite a few places that
use std::cout (for one example) directly (e.g. HashS::showhex and
Slicer::SeedMe).

IMO, these should take a stream as a parameter, and use that stream,
rather than always using std::cout and std::cin.

As-is, your HashS class looks to me like it's a function in hiding (so
to speak). File_spec looks a bit similar -- you basically just
construct a File_spec from a string, and then use (public, no less) data
members of the object. I see little advantage to this design.

It's also somewhat confusing that encrypt.cpp contains only a driver,
and the encryption proper is done in slicer_class -- the latter name, in
particular, gives not even a hint as to the real use or point of the
class.

Though I've been _trying_ to stick to commenting on how the code is
written, I can't help pointing out that your design for the file format
makes a ciphertext-only attack _dramatically_ easier than it should be.
To get any hope of security, you want to make it as difficult as
possible to even guess at whether a decryption was successful or not --
if at all possible, you'd like almost any key to produce something
that's reasonable, so it's as difficult as possible for the attacker to
decide whether a given key is correct. Your design does the opposite:
it tells him immediately whether his guess at a key was correct.

 

[Martijn Lievaart]

[ ... ]

I must be pretty dense...I had the hardest time understanding what I
was doing wrong. I think I get it now. Will you take a look at the
(revised) code from my first link in this thread and tell me if my use
of headers is now more or less normal?

If I still had the link handy I would, but I don't...

http://home.bellsouth.net/p/PWP-brightwave

Sorry I didn't include the link before...I didn't want people to think
that I was entheusiastically pimping my novice code, and the link was
just up the thread, at least on my newsreader ;-) Anyway, thanks for
responding.

g++ -c -o J_Crypt.o J_Crypt.cpp
In file included from J_Crypt.h:9,
from J_Crypt.cpp:1:
hashSclass.h:6:30: alignint_class.h: No such file or directory
In file included from J_Crypt.cpp:1:
J_Crypt.h:10:30: filespec_class.h: No such file or directory
J_Crypt.h:15:23: encrypt.h: No such file or directory
In file included from J_Crypt.h:16,
from J_Crypt.cpp:1:
unencrypt.h:9:26: HashSclass.h: No such file or directory
unencrypt.h:13:30: filespec_class.h: No such file or directory
In file included from J_Crypt.h:16,
from J_Crypt.cpp:1:
unencrypt.h:15: `Filespec' was not declared in this scope

(snip a lot more errors due to missing headers)

make: *** [J_Crypt.o] Error 1

Compilation exited abnormally with code 2 at Fri Nov 7 10:18:11

It really is better to watch your case.

A quick look on some random chosen files.

- hashSclass.h
You only need <string>, move all the other #includes to the C++ files.

- J_Crypt.h
A file that includes all other headers. A style issue, but I don't like
it. Also, including <iostream> may lead to code bloat and longer
compilation times. Include it only where needed.

- Encrypt.cpp
Looks good. Good comments, very readable. Only thing is that you do user
interaction here. Better to decouple this. A way to do that would be to
pass a const ostream& or a function pointer to a function that can write
strings. The latter technique makes sure you can use the same encrypt
routine in a GUI program as well.

- menu.cpp
* Get text by line and parse that, good!
* You may want to use a std::stringstream to parse the integer, else just
use atoi.
* Why not print the menutext the first time from menu1 as well?
Inconsistent.
* You may want to flush the output stream after printing menutext, in
trhis function you cannot be sure it is terminated with a newline.

HTH
M4

 

[Martijn Lievaart]

- J_Crypt.h
A file that includes all other headers. A style issue, but I don't like
it. Also, including <iostream> may lead to code bloat and longer
compilation times. Include it only where needed.

Forgot to say: Use <iosfwd> in headers instead.

HTH,
M4