Asking for software, get libraries in answers

Question

Is it acceptable to answer questions asking for software with libraries that will do most of what the OP is looking for, but would require some code to be written to utilize?

The OP can of course request only compiled/works-out-of-the-box software be considered, but what about in those cases where this is not specified? Is it reasonable to recommend libraries that will do 90% of what the OP needs and only require minimal code to be written to 'tie it all together?'

Answer 1

When it's not explicitly specified, you must assume a "simple user" asking the question. For him, that would be a clear No-Go: you cannot expect any truck driver being able to code (just an example, no offense to truck drivers!).

If unsure, you can always use a comment on the question to ask for clarification. The OP then can update the question accordingly, to either in- or exclude libraries.

As long as it's not clear if it would be OK to the OP, the answer IMHO is a clear "No".

Answer 2

I would say YES

with a few caveats:

• PRIMARY MUST: from tone of question or context it appears that the asker would consider that a reasonable solution for them (ie 'As a programmer I need...', 'I don't mind fairly complicated setup requirements' etc.) OR confirm via comments.
• MUST: include that you've used it for something similar and pros/cons
• MUST: include EITHER:
  • quick and dirty example code (or pseudo-code) to achieve what the user wants (ie due to not knowing 100% of the user data/requirements, you don't have to write it full out)
  • OR: link to instructions on how to do similar function (i.e. to programmers.SO question etc.)
• SHOULD: mention cost and license

Answer 3

TLDR: no

Why not (theory)

Software that is shipped is

• tested
• documented

If you need to write the program yourself, there are typically lots of caveats, which make the solution extraordinarily expensive. My opinion is that people come here and ask for software to reduce cost.

Even if someone recommends a library which is tested and documented, you

• need to understand the library (you need training, which needs time)
• can use the library in a wrong way (you have bugs, fixing them needs time)
• can make wrong assumptions on the input
• need to implement the product (which needs time)
• need to implement unit tests (which needs time)
• need to test the final solution (on various OS, with certain data -> time)
• need to document the final solution (even if it's only internally -> time)

This will cause a lot of effort, which certainly contradicts the idea of software recommendations (if not asking for programming languages or libraries explicitly).

Why not (practice)

In this question, the first answer was not a full product, but a programming language + library suggestion.

It was implemented (~200 hours) including some unit tests (not many) and some automated system tests (even less), tested by 6 people doing beta testing (exact effort unknown) and documented (~40 hours).

Given an hourly rate of 50 €/h, the total cost was roughly 15.000 €. Would you ever pay that amount of money for such a simple step by step test tool? Certainly not.

The first answer focused on being free, but that refered to the cost of the programming language and libraries. Being free does not mean gratis. And even if the question was not tagged gratis, such a solution certainly exeeds the budget, so it's not an acceptable answer.