|
All,
I've been unable to find a decent article on exception management strategies. I'd like to know what everyone thinks of this idea (this way, if the idea sucks, only a few people ridicule me.  )
Basically, I'm looking for a strategy that meets the following requirements:
Goals
1. Provides an Intelligible message to the user that
a) Explains the problem
b) Provides a list of possible explanations
c) Suggests steps to remedy the issue
d) Can be told what actions are allowable in the wake of an error (continue, quit, retry)
2. Is fully configurable via XML
3. Doesn't clutter your code
4. Provides backward compatibility so if a component developed with this strategy is used by an application that does not use it, the new author does not have to handle things differently.
5. Information must be preserved as the exception is passed up the stack. (more on this later)
High level design considerations
1. My thinking is that exception handling does not need to be very fast, it is more important to be very accurate.
2. Display should be performed at the highest level (usually the display level, but not necessarily), and never anywhere lower. (Displaying an error in a utility function ties the backend code to a display technology and method, reducing it's ability to be reused. It also means that the same error may result in multiple messages to the user).
Tentative Implementation Plan
Currently I'm planning to create an exception factory, and XML files for each assembly. The user might throw an exception thusly:
try<br />
something that breaks<br />
catch ex as exception<br />
throw ExceptionFactory.createException(me, ex, args)<br />
end try
The factory first would check a config file to see if it should use this new exception handling mechanism, or just the built-in .net error handling. If regular handling, then it simply returns the original exception. If enhanced error handling is used, then the factory would use reflection to determine where the exception originated (the 'me' object). Then it will hit that assembly's associated XML file to find the info on this error. In this way, the code can be used for any project w/out change. It also means that the XML file could be used as an appendix of sorts. Finally, storing this info in the XML file means that it doesn't have to be in your source code. The args may be used to pass extra information. For example, if it's a file not found error, it would be helpful to know what the path is of the file that didn't exist is. All of this info is stored in a custom exception class.
When the exception is formed and thrown, it will proceed up the stack. If there is another intermediate function, the factory method is again invoked. However the factory will check to see if it is of the same type as our custom exception. If so, information can be added to the messages, and possible actions can be reduced as necessary.
This brings me to a more detailed explanation on #5. My concern is that creating an error message way down in some utility function may not make any sense at all to a user because they may not understand why that utility function was even called. (Half the time I don't even understand!) Further, a utility function is probably so technical, that no possible error message would make sense anyway. If you're performing an autosave that fails, you want to tell the user that the autosave failed, not that the filestream could not be opened... So basically, care must be taken so that information is ADDED as the exception passes up the chain to get a message like this:
-------------------------------------------------------------------------
Page copy failed <-- GUI level
because <-- added automatically
the data could not be obtained <-- next level down
because <-- added automatically
deserialization failed because <-- added by the function that threw
the file can not be found the exception in the first place
If the file: {0} exists on a network share, please check your connection to the network and try again. If the file is gone, then the data has been lost. You may either continue without this info, or quit.
<<retry>> <<continue>> <<quit>>
--------------------------------------------------------------------------
What I'm looking for is:
1. What do you think of my requirements?
2. Is this the best strategy to go with?
3. Any other general suggestions?
|
|
|
|
|
I was thinking of submitting an article about advanced software protection. The article I saw a long time ago about software protectiong that was entited "Software Protection 2.0" did not include many of the techniques I wanted to discuss.
The article would get into some of the methods that anti-cracking software encryptors/protectors/packers use to protect other applications like shareware ect. I was wondering if this would be an ok type of a post u know it would be getting into lots of the assembly code protectors, and revealing some of there secrets such as protectors like Aspack, HASP, ect use to detect debuggers, confuse disassemblers, stop applications like ProcDump, WinDiasm ect. I was wondering if you thought it would be a good idea or not?
- LiquidKnight
|
|
|
|
|
|
I was thinking of including my PE File Encryptor I wrote in masm32 :p
- LiquidKnight
|
|
|
|
|
True. Similarly I can't protect my house from a determined burgalar but I still set the alarm and lock the door. At least that stops the casual thief.
--Colin Mackay--
EuroCPian Spring 2004 Get Together[^]
|
|
|
|
|
|
your messing up the word protect protect doesnt mean fail safe.
|
|
|
|
|
How am I messing up the word "protect"? messing up with? itself? If it looks like it was a typo...
I'm with Colin - there's no perfect protection, still, people lock their front doors.
Flirt harder, I'm a coder.
mlog || Agile Programming | doxygen
|
|
|
|
|
People lock there front doors in hope that there alarm goes off or they hear someone trying to get in. Plus, it keeps out idiots; there never has been a perfect protection on anything thats not the purpose of the word protection protection means to protect, not to make immune. Its unfit to compare a computer program to a door of a house they are entirely differn't in so many ways. Partly because most people don't walk around going up to houses and trying to open them cuz they know if they do they will end up in jail, where as a peice of software is differn't. so yea..
- LiquidKnight
|
|
|
|
|
People I know don't go around trying to break the protection on software because they know that if they do and get caught they will go to jail.
There are enough people who do go around houses and try the door to see if it is unlocked. Sometimes the person is more determined. When I was at school my parents house was broken into. I was the first to arrive home and I noticed something was odd. The patio doors were slanted, the back door had marks in it and the kitchen window was ajar. The house was a mess. When the police did their investigation the opinion was: The burgalar had arrived and trying to open the back door with a chisle like instrument, that failed and tried the patio door and the kitchen window all these attempts failed. Then he climbed onto the roof and in through the skylight (so he must have been thin). He took what he wanted then tried to get out. He attempted to open the patio doors but failed (he succeeded in breaking the mechanism such that it was hanging slighly-off its hinges but wouldn't open futher than a few centimeters - he also left a big Addidas imprinted shoe mark on the glass), he failed to get out the back door as similar marks were found on the inside, but succeeded in cracking the glass. He did eventually manage to get out of the kitchen window.
--Colin Mackay--
EuroCPian Spring 2004 Get Together[^]
|
|
|
|
|
hi liquidknight...
i tend to agree with you...
Sun Tzu (the author of Zen and the Art of War) i think said it best...
"You can be sure of succeeding in your attacks if you only attack places which are undefended. You can ensure the safety of your defence if you only hold positions that cannot be attacked."
the basic premise, however, between attempting to enter a house or a computer system is basically the same, and i say this with some authority having spent some years as both a physical security analyst and an information security anaylst... i suppose the difference lies in the fact that people feel more 'secure' hacking at a computer/file/etc as 'the nosey neighbour' can't see them doing it... but break and enter/theft/criminal damage is the same for both a house and a computer assuming the intent is the same, to deprive the owner of an object (physical or electronic) by means of removing permanently or damaging the object to a point it is no longer usable, and to enter an area (physical or electronic) for which you have not been given the authority to do so...
here in australia, it is covered (i believe) by the Commonwealth Crimes Act 1914 Section 7a which includes any unathorised entry be it electronic or physical...
ultimately, i also believe that protection can only protect from the vast majority of people who are either not interested in attacking a target (like a house or computer system) or the inept... for those truly determined, they will eventually find a way...
so, yes, you are 100% correct (in my mind) when you talk about protection being to protect and not immune, and also 100% correct (in my mind again) when you talk about there never having been any system that can protect 100%, however, i must disgree that it is "unfit to compare a computer program to a door of a house", they, from the purpose of law and general intent, are identical, it is only the motive and manner which are different...
either way, your submission idea seems good... i am looking forward to reading it...
cheers
nik
Nik Vogiatzis
PhD Candidate: University of South Australia
+++++++++++++++++++++++++++
Developing new generation Traffic Micro-simulation Tools for Traffic Engineers
em: nikolaos.vogiatzis@unisa.edu.au
|
|
|
|
|
The only problem with your analogy is that the burglar seldom picks up your house and takes it to a place unknown to you, where he or she can break into it, out of sight of anyone. 
--
Seraphim Shock. Gold for your ears.
|
|
|
|
|
i'll concede that point... 
Nik Vogiatzis
PhD Candidate: University of South Australia
+++++++++++++++++++++++++++
Developing new generation Traffic Micro-simulation Tools for Traffic Engineers
em: nikolaos.vogiatzis@unisa.edu.au
|
|
|
|
|
have a problem and I feel for most of the students it is trivial but I have not been able to have a satisfactory solution to it despite putting considerable time . (Due to its trivial nature I preferred the anonymous option/ mailing you directly that you mentioned in the class ).
PROBLEM:
I compile a simple program using visual c++
#include<iostream.h>
void main()
{
cout<<"CANADA";
}
If I copy its execuatable to linux or try to run in shell(it does not run or outputs some garbage with error message- I mean native code is not portable),
I know the different platform will not run the exe complied at windows platform and vice versa.
My problem is I want to know the reasons/factors that contribute to the difference, (OS system calls etc , compiler issue I am not sure).
Any details or webPointers/explanations/chaptersInBook on this issue (why exe for one platform does not run on other platforms though the library calls are the same- issues involved etc ) will be of great help to me.
Thankyou very much for your time.
|
|
|
|
|
|
Hi all,
Our firm is in search for the standart/sample on documenting the program product. I mean technical documentation, such as module descriptions, functions, variables etc.
Can anyone recommend some links ?
Thanks in advance,
Valery
P.S. It is late to drink mineral water when the liver fall off, I understand
|
|
|
|
|
|
Hello
I need breif documentation about peer to peer file transfer. so i can learn complete about p2p from that doc if any one have then please send it to me at ryu_fs@hotmail.com
Thnx
Foky
|
|
|
|
|
Hi, guys.
I need to develop an application which needs to comunicate through a modem. It should be able - through others - to play a wave file stored on the computer through the modem and I have no idea of how to do it. Any explication, either accompanied of code or not, might help me pass this problem.
Also, the application needs to recognize and generate DTMF frequencies.
I'll be glad if you might help me!
Thanks a lot!
|
|
|
|
|
There are several articles at CodeProject which discuss serial communications; these will help you learn to code for modem comms. Recognizing and generating DTMF tones is a hardware task performed by the modem itself. These tones are not encoded and passed through to the application software for your use. If your goal is to send .wav files to other computers, a simple modem program will suffice, but if you intend to play the audio over the phone lines you may have to design a new hardware device for that purpose. Try posting this on a programming forum and see what develops.
"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.
|
|
|
|
|
Such Applications are Developed using TAPI (Telephony API)
Chris Sells once wrote wonderfull classes to minimize the programming effort required
try seaching at google
(The Source came with a book 'Windows Telephony Programming'
i think)
C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg
|
|
|
|
|
 Hi,
First, it's a very good example.
Second, I would like to know if I can present a Column 3d graph with this example and if I can, how can I do it?
Regards, Michal.
|
|
|
|
|
Anyone interested in an article on Performance Monitor Counters?
|
|
|
|
|
Yes, as a matter of fact, I am.
Here is the dilemma.
I want to know if a process is running on the local or a remote machine.
I use the PDH helper library to enumerate the 'Process' object list.
However, this text 'Process' is not the same on foreign language systems!
In French, you need to use 'Processus'. Probably something else in German, Italian, etc.
So, using perhaps remote registry connection, I want to discover what NAME to use for the process list so I can turn around and pass the correct text into the PDH functions. I am pretty sure the correct text is stored in the dynamic data hive of the registry, where the performance data is stored.
Can you perhaps provide a sample of enumerating or traversing the registry performance counters and interpreting the names of the 'objects' that can be passed to the PDH functions?
|
|
|
|
|
hello can somebody help me?? i just wish somebody that help me to give me some idea about some project or system that is relate to the networking title so that i can have this idea to helping to enhance my project
so so thanks for you helping. thanks
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
