#Market Survey
16 messages in this thread
Perry, I'll accept a truce, for now.
Consider this however, you sell me a game, I register it, you give me the
number which unlocks the game, I input the number and trace the execution to
find the lock, change the lock and number to a random combination, make 50,000
copies and distribute them. I don't think volume pirates will mind this at
all.
Casual pirates, friends who get together and buy one copy and dupe for the
others will love your system. It does not effect them at all.
Do you begin to see why this is unproductive, senseless and if anything,
generates bad publicity for you?
Ted
Ted,
I don't think the lock is the key in Perry's form of Purchase
Verification. All he needs is the 5 digit serial number he has imbedded in his
code, somewhere, from the pirated copy, and the registration card that
corresponds to that serial number to be able to point at the last person that
legally owned the software <<– The Pirate.
Since the hidden serial number will never be accessed by the code
itself (Perry is the only one who may know what he is loking for), then I think
it would be pretty tough to trace.
-John
John – there is no way to hide such an imbedded trace; that would simply be
removed as well. Don't ask me how, as I'm damned if I'll publicly reveal how to
destroy someones livelyhood, but rest assured that you _cannot_ do it
effectively. I would estimate the time required, for someone who knows how (me,
for instance) to find such a trace at about 1 minute per 100k of program. You
want sales that justify a product? make it a fabulous product. That's the best
chance any of us have. Cp sucks. –Ben–
Ben,
Im beginning to see the errors in my line of thinking. Copy Protection
in any form that I know of can be beaten. Purchase Verification may work at
some point, but it would take more laws. Who wants more laws. Im going to have
to spend some time and rethink my ideas on copy protection and verification.
-John
Well, don't feel bad. It's a problem without a solution, and there have been
many minds at work on it before us – not one has come up with a reasonable
solution that doesn't require significant overhead by the software producer…
And the hardware manufacturers won't co-operate, either. Better programs, more
profits. –Ben–
Ben,
I thought PCLO has an embedded trace. Why do you bother, if it can be found
and removed so quickly?
PCLO is _not_ an imbedded trace – that is an assumption that was made here, and
I allowed to foster itself as it seemed to be doing me a favor – The method
used in PCLO is similar to an imbedded trace, and that's why/how the assumption
was generated. However, if you were to examine two copies of PCLO, you would
find (at least if they are the same revision) that the only differences in the
code are the serial numbers, which you can remove very easily with tools such
as NewZAP. I'll tell you what PCLO does – because I feel that it too will be
discovered in any case, and that our customers have been so good to us (Quality
made it so, I hope) that the method of protection has become irrelevant. PCLO
encodes the files that it creates – try saving the same file twice and see what
happens; they are, although the same when loaded into the workspace,
_completely_ different when stored to disk. The files are encrypted – and the
key contains a doubly indirect reference ti (to) the source of the program, as
well as the information needed to decipher the file records. I will leave as an
intellectual exercise (a minor one) the answer to the question of where the key
is retrieved from – for, as you'll soon realize, somewhere, something _has_ to
be different from copy to copy in order for the encryption to be unique. While
I'm at it, here's a _foolpoof_ method of identifying software – but it's effort
intensive. Anyone who wants to take advantadge of this may feel free to do so;
In software where the final executable file is made up of many object modules,
one uses a rather simple proceedure to force a linker to put them together in a
different order for every copy you ship. Once these are resolved, it's
virtually impossible to change more than a little. For program with, say, 10
modules, that creates an _awful_ lot of permutations. PCLO, for instance, is
comprised of over 20 distict modules. I still say that the only answer to the
software piracy problem is quality software. If it's junk, not only will they
steal it, they won't even _use_ it!
Guess I'll read the thread through the next time. C-: I'm giving
up on the diskettes and investing in better manuals.
Well, then let me do it. You buy two copies of the program and thereby have
two different imbedded codes or strings. Run a byte by byte comparison program
on the two files. This will expose the exact point of difference.
There, now wasn't that easy. Now we don't have to worry about that anymore.
It's exposed for the simplicity it is. Of course there are more involved
schemes, but you get the idea.
I don't like silly little secrets!
Ted
Ted,
Personally, I'd embed my five byte ID amongst 100K of noise, and I'd change
the noise on every disk, too.
OK, fine. But what it every executable out there was different? I wrote a copy
protect scheme like this for a company once. It wasn't hard to do – just wrote
a program to write link files, then the copying process included a separate
link step to produce a unique executable. Trust me, anyone with the source
code in hand will have a definite advantage over some Joe with a debugger and
200K of binary data.
Well, ted, odds are you've put it in the hands of at least one individual who
will use it to hide his theiving. I wonder if you think the design of an atom
bomb should be public knowledge, or a potent knock-out solution that leaves the
memory jumbled, or the access codes to the computer at the bank that holds your
money, assuming that you use a bank. Even in a free society, some information
(in my opinion, at least) deserves at the very least, discretion in it's
dissemination. As far as getting the idea, sure. I get it. –Ben–
Nah, but then the design of the atom bomb is public knowledge. Pick up a good
encyclopedia or tech journal. I just prefer NOT to keep silly little secrets.
All you did was tease the people who are just getting into software. I
appologise for the excessive sarcasm, but then I have a good teacher. <Just
kidding old buddy>. The heat of this discussion is peeling my wallpaper. But,
that's what a controversial topic will do for you.
I'm leaving the thread now, as I've discussed this with Perry in co. Any
comments directed my way will be answered by E-Mail.
Ted
If you use organization the code (IE: link 10 different modules
in different orders to create 10! serial numbers), you could create
code which could be uniquely identified. Unique identification
doesn't necessarily buy you much since a single stolen diskette
could be duplicated at will without any way to find the originator.
Also it would be prohibitively expensive to create unique diskettes
for medium-volume sales (100's to about 10K units, where most
Amiga programs would fall).