I know these are idle questions, but I'll try to give them a serious answer because I have some time to kill while I'm installing this new, er..., ahem...

1) Inkjet printers now come with enormous bloat to help sell the vendors other products. They can give away a printer if you just keep on buying their inks, so they want to give you an application that measures the ink levels. Then there's other printer applications, like scanning, copying, networking.

In short, the small price of a printer is subsidized by this bloat.

You probably can just plug in the printer and let the system install it with whatever it finds and it will work just fine.

2) The printer spooling subsystem is enormously complicated. A printer is now part of the network and has it's own accounts and resource managers. It doesn't like to be stopped or interfered with and it takes a while to resynchronize resources.

3) Killing a process involved killing perhaps many threads. Each thread has a life of it's own and sometimes has to wait for other things to timeout. Then things have to be flushed to disk and the disk logs flushed, etc.

4) A reboot is required after an install because some of the files are in use when a program is uninstalled. While they are in use they can't be deleted. The restart is needed to delete the files when they are no longer needed. Then the new files get used for the first time after that restart, so a second restart may be needed to start up the application with all its new-new files.

5) I don't understand that question

Single user systems back in the good old days of DOS had applications that could totally control every part of the system. Now everything is shared and there are countless layers from GUI to kernel and several privilege rings in the CPU execution unit.

Here is an interest experiment. Look up Process Monitor from Sysinternals Suite and run it then load a simple application like Notepad. Open a file. Look at the stream of stuff in Procmon that has to be done just to find the file directory.

Or look at an old web site source code and then something from Google... Things are just getting extremely complicated so that they can appear to be simple for us dummies. No one should have to learn assembly debugging anymore, just as no one should have to know how to take apart a car.

Look at the simple LED lights now - they have a class D switching supply in them that can run on just about any voltage and current source in the world. This simplicity of operation requires enormous engineering complexity.

I don't install any printer drivers.  Of course, I have a small monk do my printing and he is pretty slow...

I had a little Monkee once.  He was a miniature Micky Dolenz, only about 3" tall.  He had all kinds of strange stories about life on the road.  One day, we were going to meet at a railroad station 4:30.  I forgot the name of the station, though.  Clark's Fill...  It was something like that. 

Never saw him again, so I'm not really sure what happened to him.

Maybe he went to work for Craigb.

Beagle


1) in this case I could not plug the printer in and install just the drivers.  I did try that and it did not find the drivers.  the printer is relatively new (less than 2 years old but still in a box) but we're running WinXP on the testers.  Remember that I work for a government contactor.  These testers are actually government owned equipment which is designed, built and maintained by the company I work for.  So we don't change things quickly or easily.  thus, we're running XP with NEW 2 year old printers.  the drivers were not already in windows, I HAD to load the drivers that came with the printer.

2)  no, as I said before it is NOT part of a network.  it is connected directly to the computer via USB cable.  there's no reason, IMO, that the printer spooler should be that complicated for a local use only.

3) so?  I'm a hardware engineer and while I do write software, I tend to think of interrupts as digital - it's either 1 or 0.  obviously windows processes are not that black and white but it doesn't make sense to me why it's not.  if the processor processes that many instructions per second - there's no reason, In MY MIND why it can't terminate those processes in under a second.  there are times when it takes MINUTES to terminate all of the threads for a process to shut down.  it's just poor management IMO to allow a process to take that long to die.

4) I might concede that one, but it's still annoying.

5) doesn't sound like my problem...

while I did not really expect any real answers from these "questions" - they still are valid questions, IMO.  software vendors tend to make things too complicated when they don't have to be.  I understand some things are complicated by nature - but the basics of killing a print job or killing a windows process shouldn't be that complicated.  but it's just MO.

an LED assembly with a switching power supply to operate at any voltage can still be turned off in a few nanoseconds.    I can't say the same for killing processes in windows or kiling print jobs on a local printer!

I do not wish to argue with you because I do agree with your sentiment.
Whether you have a printer on a network or not, it is still designed as if it were on a network. You can be on a network and have no other connections (or wires).


Computer engineering is still in its infancy, and these things will get simpler (and more complex at the same time).

These days, things are usually written in tiers so that the budget-minded people (the vast majority) have something to spend and then the rest of us can add our levels to it.

That type of marketing really messes up a decent, clean software architecture. And then add licensing, anti-piracy drek and you get some more levels of interference.

And then we get into installers - don't get me started on that. No one knows how to install things properly anymore.

Just look at Google stuff - I have a Google Update service (two of them) - I have 2 Google update tasks - and I have a Google update startup program. They all do the same thing - check for updates - but they were installed in different ways by different pieces of Google - Chrome, Picasa, Google Drive, whatever... and they are at different revision levels... The different departments of Google don't seem to talk to each other.