OT: What should I do with my self-written software when it no longer runs?
 

I wrote a lot of programs back in the day, on both Amiga and PC, and as of now, all of the older PC programs (DOS, some of it) no longer runs on Windows 10, and I wonder if it's worth keeping it.

A lot of it is old graphics and cataloguing systems (one of which uses old 8.3 filenames and was my uni project!), and they no longer run and are certainly not commercially viable, and some of them were moments of madness anyway (arithmancy calculator, anyone?)

Ironically, my old Amiga software runs fine on emulation, and I'm sure I could get the MS-DOS stuff to run on DOSBox, so I might keep some of it, and my modern-day software still runs fine, so that will stay.

I'm thinking that I should at least keep the source code (for the PC stuff) even if I have no idea what's going on in it now, it shows my coding proficiency, doesn't it?

I still have my old code around, even though in many cases, I don't even know how I could compile them and can't be bothered to investigate. I found some old Amiga code I had written in 1986 or so. I also can't compile it, because I lost the second of my three dev environment discs. Still I like to keep them, and I even look sometimes at it, and there are some memories connected to it. :)
I was even surprised to see that even back then, I already adopted some good coding conventions I still maintain today, because I would have expected that the code is really bad and realized that even from todays point of view, it was not so bad after all. :D

I forgot to mention that in 2012, before the Windows versions refused to run them outright, I decided to chronicle all of my old software in a webpage, complete with pictures and descriptions, so I think that would be sufficient, as it would give an impression of how the software worked before it didn't.

I'd say keep it.
Furthermore, keep the binaries as well as the source - if you get all nostalgic at some point in the future and want to look back at it, you can bet there'll be some crucial piece of toolchain you can't track down any more!

I lost a lot of the old dos stuff. Not that it is useful, I just wish I could look back at it because it was some of the earliest stuff I ever wrote. I would definitely say keep it, for nostalgia.

My problem is that as soon as I decide I don't need something anymore and throw it out, with near 100% certainty I end up needing it within days or weeks of getting rid of it. It could have sat unused for years or even decades, but sure enough, as soon as it's gone I end up needing it. :banghead



So my recommendation is to always err on the side of preservation. Keep the code, keep the binaries. You never know when you might need them. I assume you're talking about megabytes worth of data, which can be stored for negligible/no cost on a variety of media.

I've decided I've got too much code which I don't like or which is just variations on a theme, so I'm going to cut the bad stuff only:

For example:

The arithmancy calculator I mentioned was when I was obsessed with Harry Potter and I was forced to go to a "back to work" scheme every day for a few months, so I kept myself busy by bringing along my Delphi 3 software and programming. I came across the subject and thought I could implement it in a program, but it was a waste of time as it's just fortune-telling bullshit.

Another example was when I developed a disk cataloguing program, and I'd realised I'd made a big mistake in the data format, so I went and wrote a program to "patch" existing data made by the old program, but no-one else used it and I fixed the error in later versions of the cataloguing program anyway.

A third example was when I wrote MS-DOS C programs for "intros" for the publishing company my father worked for, which were all fancy routines morphing from one image to another, or things like that. The irony was, the people at the company just couldn't get the programs to work at all... until months later when they isolated a PC from their network and were finally able to get it to work. I kept the stuff as part of my portfolio, but my dad said not to bother. I can't even remember how the whole scenario came about anyway.

On a separate note, I have written three fractal landscape generators and three "tweening" programs: think a square graphic with different RGB values in each corner, and the program fills in the space inbetween. When the RGB values pulsate with different frequency sine waves and it all happens in real-time, as my last tweener did, it can be quite mesmerising. The problem is, each set of three programs is variations on the same theme, so I'm going to ditch one of each and just keep the ones I'm most proud of.

If you want to cut it from your professional portfolio, by all means do so. But hold onto a copy for yourself. It’s something you created and has some level of value for that reason alone, even if it’s “useless”.

I do a lot of writing and a lot of what I’ve written is awful and I wouldn’t dare publish it anywhere. But I could never trash any of it. I learned something with each piece and being able to refer back to those pieces to remember the lessons is helpful. And every now and then I can steal a sentence or two for another project. :)

Once it’s gone, it’s gone...

Some of the stuff I created on PC I'd already done before, better, on Amiga. And in any case, the code I'm most proud of is in the stuff I'm gonna keep.

I'll even keep my old university project, the cataloguing system that was written in Visual Basic, runs in Windows 95 yet doesn't use the long filename format, so it was obsolete before it was even completed. But I like the elaborate lengths I went to to get it done.

As for the "variations on a theme"... Let's see, the fractal landscape generators as an example:

The first one is only 16 colours in hi-res, renders only very basic blue, yellow and green "tropical islands" but takes absolutely ages to draw them. It was basically to see how my Amiga fractal landscape code could be adapted to basic PC graphics (VGA was not available to me then) and to be honest, even the Amiga original had shadows, whereas this doesn't (not enough colours, see?)

The second one, possibly my first C program in MS-DOS and one I'm very proud of, is 256 colour VGA, has various colours for rock, foliage and water, whose elevations can be adjusted in real-time (more or less) complete with shadows cast by the sun and has an optional isometric 3D view.

The third one is written in Delphi and is fully Windows, with 24-bit colour for the 2D and 3D rendering, and instead of an isometric view, is full first-person view, complete with sun and shadow effects, and the colours are fully customisable. So good that someone used it to make their DeviantArt pictures for a while.

Guess which one I'm ditching? :)

I've re-organised my collection to feature my favourite or significant programs only, even if some of them no longer work. I'm surprised how little space they take up, they could be attached to emails, assuming the clients don't kick up a fuss about "malicious code detected".

EDIT: Yep, just tried sending my entire software library via attachment to my two email clients: Yahoo accepted it fine, while Gmail complained. Typical.

Maybe even upload to archive.org? Especially if the source code is there, you may find someone bumps into your work and takes it further or uses it as a way to learn.

Can anyone upload their stuff to archive.org? Is it free? I thought it was only for websites. I mentioned a website, but mine is more "offline".

I don't know how to do it to be honest, but I have downloaded software and books from there. On some of the podcasts I listen to they talk about uploading scans of retro mags and software to archive.org.

They probably paid for the service, I'm happy to use my standard free cloud storage. The website attachment was rejected by Gmail as well, but I've still got Yahoo to store them on, as many copies as I like, as well as local devices including two 250Gb SSDs in external enclosures.

If you're happy to share the source stick it on github.
I've been interviewing programmers recently and this is a good sign.

A good programmer should exhaustively document every single line of their code, and I don't do that now, I prefer to let the code do the talking.

The problem is, looking back at this old code of mine, I have absolutely no idea what it is doing now, but obviously I did back then.

Besides, I don't think I'd do programming as a job, but just as a hobby.

I'd say, keep everything.

I have kept all the software from the late 90's when I started PC'ing, they take virtually no space compared to anything from today so at least it's not a storage capacity issue. Every now and then I clean up my HD's and wonder will I ever do anything with ncd.exe (Norton Change Directory, I believe) etc but I have still kept everything.

AND JUST TWO DAYS AGO! I came up with something for what I needed *REAL* MS-DOS 6.22 / Win98SE hardware and all the stuff stored for years were there just to copy and use :-)

So, keep it!

Too late.

I kept all the better stuff I'm more proud of, yes.

Old DOS programs take up very little space compared to newer stuff. I have an old DOSBOX folder of stuff I wrote and used back in the day. Even with the install of TurboPascal and TurboC that I worte it with I'm still less than 100Mib.

What a load of rubbish. The code should be self documenting. If you need to 'exhaustively' document every line you are obviously writing bad code.