What is the most infeasible application/program you've ever been asked to write?

Question

I hope, certain times people may ask you to write strange things.

What is the most infeasible application/program you've ever been asked to write?

Answer 1

A few years back we wrote an Online Task System (Fairly simple ASP.NET app w/ SQL Backend) It was used for some fairly simple bug tracking/issues management for internal projects.

We got a requirement from business, that they wanted their outlook email to appear in a panel below the bug list...

(Essentially the wanted to integrate their local outlook PST file into a panel in the browser window).

Now I'm sure that it was probably possible through some absoultely horrendous activeX implementation or something but we just said "No"

Answer 2

I was approached by someone to write a replacement for Windows (this was right before Windows 95 came out.) They claimed to have half a million dollars to spend on it.

I didn't need to think long before turning it down.

Answer 3

I was once working for a company that decided to write their own XML parsing code because they felt they knew better than Microsoft or anyone else. I was asked to contribute to that team, but refused on the grounds that I thought it was a mental decision.

It turns out that most of the bugs in their app were around parsing XML that they hadn't considered, and in the next release they retro-fitted MSXML.

Answer 4

Years ago, I had to work on a application that scanned recorded phone calls from a prison facility phone system, looking for particular keywords or phrases uttered, and flag the recording where possible matches were found so that prison staff could be alerted to investigate the recordings further. The idea was that if an inmate said anything suspicious ("jail break", "kill", "knife", etc), the staff could be more prepared to take any necessary action.

The voice recognition software was already provided in a library format, so my task was to write the code to pick up the recordings, scan them, and log the keywords.

There were several infeasible things at work here: Voice recognition technology was still fairly new at the time and not very reliable (and judging from my cell phone, it still isn't that good). Voice recognition in general is aimed toward more interactive applications anyway, and does not fit well with phone recordings or even live inmate call scanning (Heh. . . "I'm sorry, did you say 'Have a good night' or 'I have a knife'? Please speak clearly!"). Plus, with inmates speaking Ebonics or other pseudo-languages, people yelling and screaming in the background, and doors slamming, there was just no way to make this software work as well as it needed.

But I went ahead with it. It was my first big project with this company, so I did my best, and gave an impressive try. I was happy when the next new developer came along and got the task of maintaining it, so that I could move on to more realistic projects.