i've found scrypt for delphi but desire to use it cross-platform and don't know enough about removing stuff for windows... is there an scrypt available for FPC or lazarus that is easily implemented and used?

scrypt

http://www.wolfgang-ehrhardt.de/crchash_en.html

Am Monday 26 October 2015 20:13:49 schrieb : > scrypt http://www.wolfgang-ehrhardt.de/crchash_en.html

the license on that one is too restrictive :(

What is so restrictive about zlib?

where do you see anything about zlib?

in the page linked first paragraph/line from the top [quote]
Here you find a few Pascal / Delphi source and binary files related to CRC, hash, and HMAC calculations. All files can be used and distributed under this license
<http://www.wolfgang-ehrhardt.de/info_en.html#License>, the original template of the zlib/libpng license can be found here: zlib-license <http://www.opensource.org/licenses/zlib-license.php>. [/quote]

ahh... but the problem is [keep reading]

[quote] Here you find a few Pascal / Delphi source and binary files related to CRC, hash, and HMAC calculations. All files can be used and distributed under this license <http://www.wolfgang-ehrhardt.de/info_en.html#License>, the original

the above link is the license URL... the stuff about patents is off putting to say the least...

template of the zlib/libpng license can be found here: zlib-license <http://www.opensource.org/licenses/zlib-license.php>. [/quote]

the above link is the license URL... the stuff about patents is off putting to say the least...

Apologies for chopping the bulk of that message, but it's coming over unformatted.

The patents issue relating to zlib etc. is way out of date, see https://en.wikipedia.org/wiki/LZW#Patents for a summary.

What you choose to use depends on what (if anything) you need to be compatible with, but for some password stuff I wrote a couple of years ago I used Tiger2 from DCPcrypt.

no problem... i do take steps to reformat some quotes some times...

The patents issue relating to zlib etc. is way out of date, see https://en.wikipedia.org/wiki/LZW#Patents for a summary.

it isn't those patent issues... there's something else that the license link page talks about...

What you choose to use depends on what (if anything) you need to be compatible with, but for some password stuff I wrote a couple of years ago I used Tiger2 from DCPcrypt.

understood on the compatibility... i was just pointing out that that implementation was not right for our needs... maybe there is another implementation that will work... if not, reinventing the wheel will be necessary...

The stuff about patents?
Those "patent stuff" as you put it have nothing to do with the code's license you have to be aware of them regardless the library you use or the license the library comes with. They do restrict your use cases but its not the library's fault or role to solve them for you. It is the same with every mp3 library out there, you need to be aware that mp3 is a patented algorithm and if you create a commercial product using it you have to pay for the privilege regardless of the library's license.

In any way I have my answer and thank you for your time.

Go on...

The Unisys and BT patents are the ones that have given most people problems, which is why I cited that page.

crypto patents or something, i guess... it says that if there are any patents in the software that the code cannot be used... click the link and read it and you'll see it, too...

The Unisys and BT patents are the ones that have given most people problems, which is why I cited that page.

yeah... AFAIK, zip or zlib stuff isn't used in this... or is it?

That's not what it its says is that even if you do use an other library, the scrypt algorithm is the same so you do need check for patents on it and comply with your countries patent laws. In any case the library's license is less restrictive than mpl.

That's standard boilerplate, which he put in specifically because there was a handful of attempted patents covering compression algorithms (Unisys, BT, PKWARE and at least one held by IBM). I think the last of those expired in something like 2007 and I'm pretty sure that zlib etc. predates that... apart from anything else any compressed GIFs were believed to sail close to the Unisys patent.

As I said, he's using the same boilerplate license for all his stuff.

I am not a lawyer, but my understanding is that software patents have not been thoroughly tested in any jurisdiction. I think the consensus is that an algorithm probably can't be patented while a specific method of using one or more algorithms might be patentable, which means that you can probably use any published hash algorithm with impunity but something more complex like Diffie-Hellman might be less clear-cut.

In any event, if the details of e.g. a hash algorithm have been formally published then it probably can't be subsequently patented, and if the author didn't reveal that he'd applied for patents before publication he'd not make himself popular. For an extreme case of this, consider the Rambus debacle.

There might be separate issues relating to copyright, but truth can't be copyrighted and an algorithm is, by definition, an expression of truth: there's no scope for opinion or subjective decisions. As an example consider SCO's attempts to claim ownership of the Linux kernel (and the row Oracle and Sun are having over Java), where the important point is whether stuff which appears in both has been copied or is quite simply doable in only one way.

So as I've said, I'm using Tiger2 for hashes since (a) it was openly published, (b) if Ross Anderson finds I've screwed something up and feels it's to his discredit he's close enough to make it worth being polite and (c) if he /really/ doesn't like something I've done we're in the same legal jurisdiction.