Printable View
I will be posting completely executable samples that will be as tamter resistant and with an internal algorithm as secure as and native application.Quote:
Originally Posted by WizBang
Additionally I will re-open the challenge of someone producing a peice of code with a given set of performance specifications that I can notmatch within 10% using pure managed code....
Original source is convenient but it isn't necessary to reproduce an application. If you have a native exe, given the motivation, you can produce the source for equivalent functionality. Is it the original code? No, but does it matter if the functionality is copied and the app operates the same way?Quote:
Originally Posted by RealityIsForLocos
What is so difficult to understand that a native app can be reverse engineered? It's a matter of effort. So .Net by default is easier to reverse engineer. So what? If is important to have a false sense of security that you have with a native app, then obfuscate the .Net code or ngen it. Happy now?Quote:
Originally Posted by RealityIsForLocos
What is this faith B.S. with regard to Microsoft? It's not like Microsoft hid this fact when 1.0 came out 6 years ago? If you are moving to a new framework, then it's up to the developer to understand the differences. It's funny to see folks post that they just discovered Reflector exists after using .Net for years and somehow blame MS.
This whole post has gotten off topic with regard to using VB6.
The folks here that are recommending moving to newer technology are suggesting it because if a developer remains up on later technologies, historically it's shown that they will remain employable. Consider how many C programming positions were available 15 years ago as compared to now.
Do you think VB6 will be any different?
Now actually I don't really care whether folks move to newer technologies. In fact I like it in my business if I'm bidding against competitors offering VB6 solutions to my clients. It's pretty easy to land those bids.
I think there needs to be an age minimum on CG. That way we don't get the 12 year old hackers here.Quote:
Originally Posted by RealityIsForLocos
In case you're wondering, CG has a Article section where people who love sharing their knowledge and experience with others. Nothing wrong with sharing ;)Quote:
Originally Posted by Joeman
This issue also always arises, and my answer to this is always that your logic and planning is wrong. Trust me, I should know ;)Quote:
Originally Posted by Joeman
If you design your .NET application well enough, it wouldn't be slow.
Let's keep this discussion civil, and leave out any personal comments, ok?
See my previous posts.Quote:
Originally Posted by Arjay
Really what is the point of continuing? The original post was a poll about using VB6 and I believe the intent was to try to find out why folks aren't using later technologies.Quote:
Originally Posted by WizBang
Then we got off on the tangent.
Is this the only reason for not using later 2005 or 2008 technologies?
Does this mean you haven't an answer to the fact that obfuscation and ngen don't secure .net code nearly as well as a native C++ executable? Once again, hacking or reverse engineering aren't even close to decompiling. There's no point in continuing to suggest that it is.Quote:
Originally Posted by Arjay
I don't think the topic got off. But the reasons not to use .net may depend on who you ask. IMO, just because something is newer does not make it better. Take winMe and Vista for example. Remember Passport? I am not one of those who jump on all the latest that might come along. It has to not only prove itself, but be of greater value to me than what I already have. Thus far .net has not demonstrated this. It is not a step in the right direction for me. I don't work in some large corporate office, writing apps for internal use.Quote:
Originally Posted by Arjay
Anyway, as I've stated, writing programs for the home PC market is quite different than for private business use. The ease at which .net code can be stolen and/or compromised makes it unsuitable for wide distribution. This is one of the reasons I suspect ms has been encouraging pay-per-use software, where the user would log in to a server and access an app via a web interface.
So I'll be interested in seeing what TheCPUWizard comes up with for protected .net executables. Anyone else who wants is also free to post some examples. Let's see some proof one way or the other.
Any one else catch this?
Please be FACTUAL in your posts. One lie, and it's all over.Quote:
I bet if you program in the .net professionally, all you be dealing with databases and management and not with the newer cooler games and ipods.
I doubt you're writing anything for IPOD's either.
Over the past 30 years, almost any software that has been produced has been hacked to some degree. Now, the debate seems to be whether VB6 was more secure than a Net app.
ANY vb6 can be 'jumped' if it uses some kind of log-in/password jump that does one thing if you are correct, and ends the program if you don't.
That is a sequence of about 8 bytes, that can be searched for and replaced (by reversing the condition). Someone wrote the vb code, and posted it somewhere. I got it before it was pulled down.
I *thought* my app was secure, but, once the sequence is re-written, the app can be given to anyone without requiring registration.
This app is about as secure as you can get. The actual .exe is not stored anywhere on the pc, and even if you hacked the version on the pc, it is overwritten the next time you run it.
http://www.siokonline.com/publix/publish.htm
this requires IE, though. It just displays you a listing of the files on your primary drive.
Well.. I just tryed the dotfuscator that comes with VS2008 on a fairly complex app that I upgraded from VB6.
The upgrade went pretty well, only had to change about 6 or 7 lines of code.
The initial build in VB 2008 loaded up fine and appeared as though the program should work.
The build done with dotfuscator crashed hard on startup.
Conclusion it did a fine job of protecting my code as no one will ever see the resulting executable.
dglienna, did you read this thread from head to toe?
Yes, not all people can nor will do this, but making hardcore games can't be done in the .net.You just don't have enough speed in the .net to do such a thing. That is why he said that and also, you do get alot of jobs with database management for the .net.Quote:
Originally Posted by dglienna
dglienna you are way off. We didn't say things couldn't be hacked. We are saying you can decompile the .net app to source code from what it originated from. Please read this thread more carefully next time.Quote:
Originally Posted by dglienna
That is sad :(Quote:
Originally Posted by WillAtwell
Well I base my posts on facts, and not theories, does not mean my post is wrong.
The .net is slow and you don't see any new games out there that are being built out of the .net.
Companies, like EA games, need to protect there source code and get all speed they can get
to stay on the top of their competition. Do you think that if EA games started programming
in the .net, they would fall? If the .net should be on a managed and secured computer, that
sounds like a server and not a regular pc users computer.
PS ipods are mac not a micorsoft product. So go and try to program for an ipod with the .net.
and by the way if you program in the .net and get a job the first year all you
will be doing is Customer Service lol
I think something interesting is going unsaid here, and that you're implying something that perhaps you don't mean to imply: Does the fact that any publicly released software can be reverse engineered mean that there's no commercial importance to keeping a project closed-source? Would it really make no difference if all applications were distributed with source code?Quote:
Originally Posted by Arjay
These are not rhetorical questions, as I am genuinely interested in their answers. Also, .NET is not the only cause for this debate... those of you who say that the source code of your C++ work cannot be gotten are obviously not writing and distributing libraries that rely on templates.
Didn't I answer with the second post here? Of course I've read it.
This is NOT about the open source debate, either.
If games get too hard to protect, then they will just release a version such as that, that runs only online. While VB6 wasn't suitable for writing games (unless you used API's) VB.Net is much different. Using GDI, it's much faster, and does things that were very hard to do with API's before.
Just compare the IDE for C# and VB. Which one is more useful? I'd say the VB editor. :)
from http://www.cplusplus.com/doc/tutorial/templates.htmlQuote:
Originally Posted by Hermit
Now I don't know if this is true, but it sounds like they split the function up to match up to what is supplied until it doesn't find anymore different types of inputs and then compiles them separately.Quote:
When the compiler encounters this call to a template function, it uses the template to automatically generate a function replacing each appearance of myType by the type passed as the actual template parameter (int in this case) and then calls it. This process is automatically performed by the compiler and is invisible to the programmer.
Can you provide resources that can counter attack this claim or explain why this would give your source code back?
You did post 2nd on this thread, but it is on page 8 nowQuote:
Originally Posted by deligeena
??????????????????????????????
I don't think you followed this thread at all. Did you even understand it? I never stated anything about open source is a good or bad thing. I stated the .net is bad if you don't want your source code to be hacked by a .net decompiler and have your source code stolen. I like open source code. That is where I got all my knowledge from.
HOW CAN YOU GET QUAKE4, HITMAN: BLOODMONEY, STAR WARS: BATTLEFRONT, DOOM 1, DOOM 2, DOOM 3, PAINKILLER, UT2004, UT 2007, TIMESHIFT, SOLDIER OF FORTUNE: PAYBACK AND ANY OTHER GAME LIKE THIS ONLINE TO BE PLAYED LOCALLY???????????????????????????????????????????
uh..................................................................................................
So now that we seem to be beyond the 'secure code' debate, it looks like the discussion has turned to revolve games and that .Net isn't suitable for writing games.
Is that what you guys are saying?
Templates are source code, and cannot be compiled in the usual sense (rather, they are instantiated into something that is compilable). If your library consists mostly of templates (i.e. if they are part of the library, not just used internally by it), those templates must be distributed as source code.Quote:
Originally Posted by Joeman
Anyway, I don't mean to stray too far off topic, though admittedly I couldn't care less about the fate of VB6 :)
Same thing that I posted. Click-Once technology, built into .Net
You click on a web-link (only while the site is up) to download the current exe that runs locally in memory.
There are online courses that are written in .Net technology. Look up InnerWorkings.
Try to remain professional, if you are able. You didn't respond on page 2.Quote:
Originally Posted by Joeman
In C++ template's can't be exposed in compiled libraries. You need to include the headers files (like the way the stl is distributed). Hermit already stated this.Quote:
Originally Posted by Joeman
Yes. Do not take my simplified explanation to mean that I don't know what I'm talking about. There are further complexities involving the export keyword, but in general, template libraries are distributed as (original or obfuscated) source code.Quote:
Originally Posted by Joeman
I would not go further than a quick mention of Microsoft's XNA framework in that argument. I'm still very interested in an answer to my questions here regarding code security.Quote:
Originally Posted by Arjay
what is a dll then? and i said like on page 2 for a reason. maybe like on page 3 then it is there
.NET is not supposed to be a framework for developing operating systems. You certainly cannot develop an OS in VB6, so I don't see the point of this argument...Quote:
Originally Posted by RealityIsForLocos
You don't appear to understand how C++ templates work.Quote:
Originally Posted by Joeman
There's a reason that the old C string routines like strcpy were distributed in a dll, with a .lib file and a header. To use, C++ devs need to include the .lib file and header in there projects.
Whereas templated classes like std::string are only distributed in a header. To use, C++ devs only need to include a header.
There's a good reason why this is so.
But we don't need to get off topic discussing C++ templates.
My mistake on that. I know you can't hide your template code from developers since it has to stay in the header file, but you can compile it with the project and give the app to the users. I wasn't thinking very clear on the topic of linkage lol. Now back to why vb6.0 and not vb.net. Now since the template code isn't a worry since it is in native code, you don't have to worry about the header file anymore
Yes I know that but I was showing that the .net is very limiting and their are better programming languages out there if you are considering using the .net.
Please refrain from personal insults.
Actually it already exists, has been shipping for about 3 months, and is installed in over 100,000 devices already.Quote:
Originally Posted by RealityIsForLocos
Now your talking.. Why is it a strong thinking that .NET is not sutable for games.. Here is a Game in VB.NET, low resources, high speed.Quote:
Originally Posted by Arjay
Yes it may be a simple game, however it's a game. And it works.. The same was done in VB6 , and soon there will be more, I've already found a few reasonable games done in .NET....
That line of thinking does not work, so far, my findings are that the graphic core code in .NET is slightly faster than VB6..
Does this include win32 API calls? I think it would be interesting to write some simple loop which calls a GDI function several times, both for VB6 and .net, and see how they compare.Quote:
Originally Posted by GremlinSA
not quite the same. with GDI you can create the graphics image, and display it when you're done. i know it can be done using double-buffering in vb6, but requires a lot more code
I'm not sure I follow you. Are you saying that the .net graphics which GremlinSA referred to aren't direct GDI API calls? Either way, It would be interesting to see which is faster - VB6 or .net.Quote:
Originally Posted by dglienna
FYI:
Quote:
Private Message: Re: Samples
Today, 12:35 PM
WizBang
Super Moderator
Power Poster Join Date: Dec 2001, Posts: 5,676
Reputation:
Re: Samples
--------------------------------------------------------------------------------
Sounds Good. I'd recommend posting this decision in the thread so everyone following it won't be left thinking you chickened out or something.
Quote:
Originally Posted by TheCPUWizard
I have not forgottren them.
However based upon further thought and some practical issues, I have decided to upgrade the "project" to a full article and some sample code.
Since this will be published, and attacks actually invited, I am altering the techniques away from my "standard" approach for self-preservation reasons (since we have all agreed that any protction scheme CAN be broken).
Hopefully it will be finished beofre I leave on summer holiday on July 16th....
As soon as i can get some time, i'll try put together proper samples in VB6 and VB.NET to test the theory.. My observations were based on the simple asteroids game from the article, I cheated in the .NET ver, I did a VB6 upgrade on the project to get it in .NET .. and then from there, i could just fix all the 'Problems' ..Quote:
Originally Posted by WizBang
I found that alot of the code was no longer needed as the graphics methods used in .NET differ greatly from the VB6 methods..
Also all the API graphics methods are now tied into the graphics object, how this affects speed is also a question..
Quote:
Originally Posted by GremlinSA
Gremmy, You have just hit on something I would have NEVER considered.
"Upgraded" applications will definately perform poorly compared to their originals in most circumstances.
This is akin to taking a "C" program and making it compile under a "C++" compiler, and then calling it a good Object Oriented Program. :sick: :thumbd:
Code needs to be architected for the target environment from the ground up. The structure of a .NET program is radically different than a VB6.0 program (or a native C++ program).
This is akin to constructing a building you using one technology (eg. Steel), but following the contruction methods of another technology (e.g. Wood).
I know exactly what your saying there.. however because 90% of the project was maths and logic's, i figured that the upgrade method was the fastest way to get that all in, there after i removed the graphics and went the .NET way ...Quote:
Originally Posted by TheCPUWizard
The logic and maths works almost exactly the same. the graphics however work almost completely differently. when testing, trying it the 'old way' i almost killed the proccessor because i was unaware that every "drawline", forces the picturebox to do a refresh. the proccessor usage went to 100% in no time. however with the better .NET method, and buffering, i reduced it to less that 10%, peek..
The next part is to do a time comparison between the two versions.
I'll see about doing that this week end.
Sounds to me you didn't enable the autodraw on the picturebox. When your done making your lines you use the picturebox.refresh. See you invoked it. It is no longer just the old way that is slow. It is the old way that is the right way that it works fast. If that isn't fast enough, use the window's api for creating a line. If that isn't fast enough, use opengl. If that isn't fast enough, your code is just plain coded wrong ;(Quote:
Originally Posted by GremlinSa
Have you done graphic work in .NET..Quote:
Originally Posted by Joeman
the picturebox control no longer has a autodraw feature...
How ever the method you described, is exactly the method used in the VB6 version...
Also in .net you do not draw directly to the picturebox object, you create a graphics object of the image / bitmap you want to draw onto, and use this object to do the drawing on ...
Gremmy..
Sounds like a buffer, which is what I always do anyway, via API.Quote:
Originally Posted by GremlinSA
Once created, there's also the DoubleBuffered option, that you can enable and use. This creates two graphics objects for you to use.
Despite the fact I've only just finished my A-Levels, I started learning programming around the same time VB6 was released, and I've been working with it happily for a good section of that time. I do subscribe to the mentality of "If it ain't broken, don't fix it" (which is why the second app I ever wrote is still running on all of my machines).
However, I'm off to study Information Systems Engineering in October, and I will be bitterly disappointed if they don't teach VB.net. Despite the fact that it has had it's issues (and the .Net framework refuses to update properly on this machine), it is at some point going to be the future.
I just wish Visual Studio wasn't so **** big. Or that I had a faster internet connection :D
I have to admit, vb.net is a LOT faster than vb6 was. I've now wired up my panel demo to a sql server express file. As fast as I can click the drop-down(s) producing 41x14 different panels. they each get drawn independently with 4 to 144 different blocks.
the old program used to draw one block on the screen at a time (using Line() statements). it would take a few seconds to draw one small panel, and about 10 seconds for the largest one.
this will be easy to upload the detached dbo to my site, and use sql server.
edit: running locally courtesy of Camtasia www.siokonline.com/blocks
Lol ..Quote:
Originally Posted by javajawa
The main reason it is so big, it contains all the languages in one ..
Compare the .NET install to having a system with all the following installed.
VB4 , VB5, VB6 (.NET 2008 has framework 2.0, 2.5 & 3.0. so it needs to be comparable)
Borland C.
VC5, VC6.
Java.
ASP.
(Hmm Did i miss any ??)
And then try running the IDE of all of them at the same time ... I dont think that any system will be able to provide enough resorces to handle all of them .. Expecially if trying to debug a app that has components written in each lang......
Unfortunatly i had the grandkids over for the weekend, so access to the VB6 system was a little limeted, and i was unable to complete the testing ...
But i will complete it soon...
Dglienna said
As a developer we try out various bits of software, and even write our ownQuote:
I have to admit, vb.net is a LOT faster than vb6 was
Typically you write some code then test to see if it works - you dont normally sit down and hammer a program you have just written for a long time.
Consequently, the experience as a developer is this (I use VB.Net)
1) Load up the program into the IDE - this typically seems to take forever
2) Click on a form / module whatever - this takes forever to load
3) Do some changes to the code, then attempt to run the application - Dot net goes into constipation mode while it tries to sort out the exe to be run
4) Finally the application comes up
5) You attempt to run something (that hasn't been run before) Dot net again excuses itself while it "just in time compiles it" then finally the program runs
6) Unless you are doing some fancy code, possibly with multi threading, there is no way it appears any quicker than its VB6 equivalent. I work with Accounting systems which really dont have much room for multi-threading applications.
THATS MY EXPERIENCE AS A DEVELOPER - Its really not much fun - Certainly not a pleasure to work with
BUT .........
Dglienna said
So where are the speed increases (apart from multi-threading applications)Quote:
I have to admit, vb.net is a LOT faster than vb6 was
EVERY DOT NET APPLICATION I HAVE HAD EXPERIENCE WITH is criticised for its bad user experience - its just so slow
Is this the fault of Dot Net or is this the fault of developers who dont know how to write FAST Dot Net Applications ?
PS: One of the most criticised is Microsoft 2008 Accounting EXpress which I presume was written by Microsoft Programmers (go to their users forum if you want an ear full)
So ............. who then knows how to write efficient Dot Net Apps ?
And ... who is going to teach us the right way to do it ?
(Good Dot Net examples are few and far between)
Well, of course the Line statement is going to be slow! Try it with win32 API calls.Quote:
Originally Posted by dglienna
I think I'll post a little drawing function, but it will be up to someone else to write the .net "equivalent". Should be interesting to compare VB6 speed using win32 API to .net graphics objects, and also to .net win32 API calls.
There will be differences between a well written program and a poorly written one for any language. So before one can say well written .net code is faster than VB6, one needs to be sure the VB6 code is also well written.
George111,
Yes VS-2005/8 takes a good amount of time to load. But typically it should only be loaded once or twice per day.
Also, it is an environment designed for "significant" projects. I just finished a job where the main solution consisted fo 61 projects and over 8000 C# source file [120MB]. Visual Studio handled this without a problem and was perfectly responsive. For example a typical build (NOT re-build) took about 20-35 seconds.
If you put together a "spec sheet" on the machine you are running, and can also capture some PerfMon logs of VS performing activities such as opening forms, making changes, etc. I would be glad to look at them (PM or e-Mail me).
A large part of my professional life is training people to write good .NET applications. (In fact Microsoft convinced me to mothball my consulting firm for almost 2 years while I worked as a Sr. AppDev Field Consultant for them in 2005-2006). I am hoping to publically post a good part of my training material soon (I was gathering it for a book deal which now looks like it is not going to materialize).
Edited: Point #4 had a blooper....
Agreed. And I would expect the following (assuming both contestants are experts in the respective environment and can write 99.9% the "best" program possible.Quote:
Originally Posted by WizBang
#1) If both contestants are constrainted to use JUST the environment (No "extern" for VB6.0, No "PInvoke" for .NET) then the .NET code will be faster.
#2) If both can use ANY available means, then the results should be visually indistinguihable (ie a person with a manual stopwatch on a sequence taking at least 30 seconds, should not be able to consistantly measure one as being faster then the other).
#3) If the program is designed to run on a multi-core machine (ALL programs these days should be), then the .NET implementation will make much better use of the available processing power.
#4) Regardless of the above, the .NET developer will have a working program before the VB.NET VB 6.0 developer. This would be event greater if the code has to be jointly developed by a 2-3 person team.
To "really" do this challenge, it would be great to get a set of teams together at a "convention" (off-site) and dedicate a (long) evening. A challenge like this probably could get sponsored....
Well, as it stands, my Dell can barely run Media Player and VC2008 at the same time. But it is on the brink...(has anyone got a computer they don't want? Or even some parts? PLEASE!?)Quote:
Originally Posted by GremlinSA
In fact, I must say most of what I'm finding after downloading 3.4Gb of software (and then installing all of about 0.8Gb or it...*sigh*) has been summed up by George1111.
However, I'm already finding huge benefits from the way the whole thing works - so I'm happy...But I'll still be interested to know which one works out faster, if you do finish those tests.
Quote:
Originally Posted by TheCPUWizard
Am I the only person who misses the time when people wrote programs to be as small and simple as possible? Or is it because I'm the only person in the world who still owns two computers which can only run win98?Quote:
Originally Posted by TheCPUWizard
The more I do the coding gig, the more I am interested in reducing the time it takes to develop. As such I spend quite a bit of time undestanding and leveraging the latest technologies.Quote:
Originally Posted by javajawa
Of course I expect reasonable performance, but 'fastest and uses the least amount of resources' really doesn't place that high on the list of priorities. I place a higher premium on how do different components interact with each other and how easily I can get the components to work together.
It seems the more experience I get, the more I leverage the features of the framework I'm working in.
Transitioning to newer framework doesn't come for free either. I've spent quite a bit of time learning new technologies throughout my career. I did this when I coded in C++ (because I took the time to learn MFC, ATL, STL, SQL and so on) and I do this now in .Net by learning the latest .Net3.0/3.5 features (WPF, WCF, and WF). I suppose I am paranoid about becoming obsolete, but I must say that every technology that I've spent time learning has so far paid off with either a simpler way of doing something, a reduction of development time or both.
It's easy to sit there and say that the newer technologies are slower - it's harder to go out there and learn the new stuff to find out if it is actually slower and if so whether the speed difference is significant.
Btw, there is a similar thread in the general forum (What languages to use) that has some performance tests between C++ and C#. The results may surprise you (and I expect the performance of VB.Net to be on par with C#).
If you have been reading my posts you know I have be at this since 1972, and also that I own a (lamost) restored 1968 DEC PDP-8. The equipement I own cost over $150K when purchased new.Quote:
Originally Posted by javajawa
Size and Speed have always been about cost (and I have posted on this at length). If you have a professional developer making $25K / yr (very good money in the early 1970's), and a 4K memory upgrade costs $15K. Then investing a few months of work to save $15K (per machine!) is a no brainer.
But look at today. A top notch developer (New York City, USA) can command a salary of $150K-$225K easily. And a 4GB memory stick costs under $100.
Using these figures, in 1972 byte of saved memory ($2.44) was worth. about 12 minutes of development time. Today, that same byte of memory is worth 3.32 microseconds of development time.
I dont know about you, but in the oldr programs, I could usually find 5 bytes in an hour of examination of a normally written program within and hour. But I dont think there is anyone who can do anything in 3.32 microseconds....