Missed chances for unique features
Last year, I wrote about creating CD Towers with practically no budget. It happened by taking the same hardware base we were using as student workstations (with IDE drives replaced by SCSI, though) and adding a bit of glue code that I wrote.
In later years, I realized I could have added a really neat feature to them. It would be something that none of the "real" CD towers which cost thousands of dollars would have. Unfortunately, I didn't realize this at the time, and as a result, it never happened. I figured I'd write about it today just to put it on the record and hopefully inspire someone else.
My CD tower design was that someone would telnet to the machine and log in as "admin". They'd enter a password we had told them, and that would drop them into a text-mode quasi-graphical interface. From there, they had a few options available: add a new cd, list what's on there already, remove an image, or log out.
Now if you think about this, they're running telnet (yes, telnet -- it was 1999) into the device, so they're not sitting in front of it. That means when they went to "add new CD" and it auto-ejected the tray, that tray was opening up somewhere potentially far away. Usually, the computer teacher who managed a given school would also take care of the tower and would have it in their classroom, but not always.
I realize now that I missed an opportunity here. There should have been a program which ran under Windows and would have used their local drive to read the images. Then it would just load it onto the tower over the network. That would save them a trip to the tower and back every time they wanted to update Kid Pix or whatever.
Besides that, this feature would have added yet another bragging point. I was already proud of the fact that my design had just a single CD-ROM drive instead of a dozen. Yes, other towers were literally piles of drives and read directly from them. Mine was at least smart enough to make an image and mount it over loopback.
Just imagine it... a tower which used the client system's CD-ROM to read discs really never needs one of its own. It could be built without that part and thus save even more money. This would make it a machine which you could truly park in a closet and then forget about it.
I never got around to it. Having to deal with Windows and raw I/O in order to make the equivalent of an "ISO" file, and then somehow wrapping all of this in a usable interface was just not my idea of fun. The users never realized that it could get any better than the telnet thing they had, so they never missed it.
Still, I know what could have been.