-2

I am attempting to parse all possible outputs of a proprietary piece of software (ELF binary) whose license explicitly forbids reverse-engineering, decompilation and disassembly.

So my question is: Could calling strings on the executable in question to list all embedded format-strings possibly be considered reverse-engineering, decompilation or disassembly?

I would doubt it, since strings is supposed to only scan the non-executable part of a binary, but wanted to ask anyways to be somewhat sure.

5
  • IMHO, the question would be out of scope only if OP would have asked if it's allowed to do it, or if it's legal. Not if just asking the technical nature of an activity (i.e. do SWE professionals consider that it's decompilation based on technical arguments and usual SW terminology) .
    – Christophe
    Jun 30, 2020 at 21:18
  • 2
    @Christophe: The question is whether or not using strings is considered reverse-engineering, decompilation or disassembly, as defined by the license. Or, in case those the license doesn't explicitly define those terms, as defined by law, or if those are not defined by law, as interpreted by courts. Whichever way you go down, "reverse-engineering, decompilation or disassembly" in this particular case, are legal terms, what their technical definition is, is completely irrelevant. 10 million software engineers can say "In my opinion strings does not constitute reverse … Jul 1, 2020 at 6:21
  • … engineering", and that won't help one bit, if there is a paragraph in some law somewhere on the planet that says "reverse engineering may include, but is not limited to, using automated tools to extract textual information from object code". Jul 1, 2020 at 6:22
  • @JörgWMittag Sorry, I could not find the wording “as defined by the licence”. As such, it seems to be your subjective perception, probably in regard of the license which is mentioned for the context. The important thing is just to set the right limit in the answer: the technical definitions and understanding of technical matters are in, but the legal interpretation that might be given ro it is out, because, as you say it depends on the jurisdiction.
    – Christophe
    Jul 1, 2020 at 6:49
  • 1
    @JörgWMittag The problem is that if these boundary questions are considered out of scope in both communities (too technical for law, too legal for swe), there’s nowhere to get some more insights on any side of the issue.
    – Christophe
    Jul 1, 2020 at 6:54

2 Answers 2

5

This question is at the border line between legal matters and technical definitions.

From a technical point of view (in scope of this site):

  • Extracting strings from an executable, is an ordinary operations related to reading files and filtering the output. But presenting the output or storing it means technically to copy the content of the file to another place
  • Reading strings is not disassembly. Disassembly means to transform the binary codes into assembly code. The string alone, without any assembly directive is not assembly code.
  • Reading strings is not decompilation. Decompilation is one step further than disassembly: it's transforming the binary codes into high level languages.
  • With reverse-engineering it is more tricky: reading strings is in principle not reverse engineering: it's just reading data. Reverse engineering means to understand how the software works and what algorithms it uses, how it communicates iwith other components, and what data structure is used. Again, simply reading strings, which are human readable data, is not in itself an analysis of the data used by the software to understand how it works.

From a legal point of view (out of scope here):

Legally, it all becomes very slippery. Usually, judges will appoint technical experts ("expert witness") to explain to non-technical folks the technical matters. But the technicical matters will then be interpreted by the lawyers. In the end, the meaning will be what a lawyer may convince the judges - who have no technical background - it is:

  • For instance, if in the strings you find out database connection strings, or if you try to interpret the strings, i.e. understand if there are any embedded commands, you have already one foot in reverse engineering, since a lawyer could easily demonstrate that you reading the string is for understanding how it works.
  • if you try to find out account names and passwords in the strings, it will be difficult to argue that you're not doing reverse engineering to circumvent security measures.
  • Some lawyers could even argue that trying to recognize the encoding of strings, is already an act of reverse-engineering. Fellow engineers might smile. Jurors might truly and in good faith believe it.
  • Moreover, the legal definitions changes depending on your jurisdiction.

Conclusion

In short, I am not a lawyer. For legal advice, consult a lawyer or a qualified legal advisor in your jusrisdiction.

Despite that from a technical point of view we consider that reading strings is not disassembly, not decompilation and not reverse engineering, there is a risk that in your jurisdiction, depending on the judge and the way your action is presented, and also depending on the real intent of your action, it might be qualified as an infringement of copyright, since in the end you do some kind of copy which might not be explicitly allowed.

Conversely, other lawyers may analyse the technical matters as being an exception to the general rule. For instance, in some countries, decompilation cannot be forbidden if it is for the sole purpose of interoperability.

1

It's not decompilation nor disassembly, maybe borders on reverse-engineering but who am I to interpret lawyer-speak?

Your argument that strings is harmless is probably invalid as strings does not necessarily limit itself to non-executable portions of the binary, this is a configuration option which can be overridden using command line options.

Anyway, since you probably won't do this in a public place with many people watching, what are the consequences you must fear, apart from your conscience conflict?

1
  • The main issue to me would have been including the gathered information in an open source project that parses the output of the proprietary software. The software doesn't allow an easy enumeration of all possible outputs, so this seemed to be the best way. Of course, you could always attempt to justify the source of your knowledge using parallel construction. Jun 30, 2020 at 19:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.