h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Why on earth has this feature never been removed or improved over the years? The time estimates computers give whenever they are doing anything, like installing a program or deleting large files for instance are utterly, utterly pointless! They jump from 23 hours, to 2 minutes to 50 minutes, then they stick on 1 second remaining for about 10 minutes...

I assume they work in a similar vein to car fuel remaining estimate gauges. The are only showing you the estimate based on the current usage. So if you are climbing a hill full throttle your fuel remaining will be less than a few seconds later when you're coasting downhill. Same I suspect goes for how hard the computer processor is working at any given moment.

Fair enough, but there must be a more accurate way to give meaningful feedback or just not bother at all? I was installing a massive GIS program the other week, and my opening paragraph is no exaggeration to how widely the time estimates varied from one second to the next!

We could get it more accurate, but it would take a fair amount of analysis of your configuration up-front, and this would add significant additional time.

Best wing it.

Inevitable xkcd link: http://xkcd.com/612/

The estimates are actually pretty stable unless you introduce a rapidly changing variable (e.g. downloading something over a wireless internet connection whilst doing other things online) or try to make it do too much at once (a single 5Gb file will fluctuate a bit, five 1Gb files will fluctuate more).

haha, that cartoon is *exactly* what I mean

I know it's a hard problem, but Windows does seem particularly bad it. I guess it's a low priority.

>>..guess it's a low priority. <<

But if it's worth doing at all, isn't it worth doing well?


Progress bars are equally suspect. Once they stop moving
I'm inclined to give up, but experience tells me the page
may suddenly and inexplicably appear even when the progress
bar is showing no progress. I even imagine that if the computer
wasn't so busy doing a poor impersonation of a Time/Management
Analyst it might better spend the energy doing what it's been
asked to do.


~jwf~

"if it's worth doing at all, isn't it worth doing well? "

That is an extremely interesting question, with an equally interesting answer, which is: "No".

Consider the options:

1. Don't do it at all. Don't do ANYTHING. Present the user with no clue whatever that what they've just asked to happen has in fact started happening. This will lead to frustration, repeated attempts to initiate the action, and possible power cycling and data loss and actual physical damage to a crashed drive platen.

2. Don't do it, but indicate that it's started at least. Now the user knows *something* is happening, but has no clue about how long it will take, whether it's hung up, whether they have time to go the toilet or the shops or whether they should sit tight because it's the work of seconds. Most of the same problems as (1).

3. Spend time calculating, in detail, how long it will take to do the task, and LOCK OUT anything else that might affect the calculation to make sure it's accurate. Fine - so long as the user isn't used to multi-tasking computers. Except now everyone expects to be able to surf the net while their file is copying. It's no use having a second-by-second accurate countdown if the computer is unusable.

Or... have a guess. The user knows something's happening, knows it hasn't *stopped* happening, can get on with something else, and has at least a clue whether they should be on the edge of their seat or settled down next door with a coffee and a bagel.

The latter is the best course, which means doing it well is very much not the optimal strategy.

Which is why these algorithms have not got, and never will get, any better.

Your conclusion was a bit more pessimistic than I am able
to embrace. I remain confident that someone will figure out
what is needed, what is wanted and how to implement it.

I recently sat watching while a picture file was downloading
and a progress bar kept whipping by saying it was scanning
for viruses. After more minutes than I care to admit I opened
Windows Picture Gallery for another file and noticed that the
downloading file was already in place. But the download window
still said scanning for viruses.

I don't pretend to understand how any of this stuff works but
remain quite certain that improvements will be made if only to
further lull us into a state of user friendly compliance. But
please feel free to remain cynical and recalcitrant; the whirled
needs grumpy pessimists to keep us on our toes.


~jwf~

"improvements will be made if only to further lull us into a state of user friendly compliance"

In case you hadn't noticed, the vast, vast majority of computer users are already in a state of compliance, usually because of ignorance. You're a perfect example. You failed to register the difference between "downloading" which clearly implies that the file isn't all here yet, and "scanning for viruses", which clearly implies that it's all here and we're checking it. These aren't complicated concepts.

As I say - these algorithms will not change, because there's nothing driving them to get any better. What has changed, and what will continue to change, is the speed of operation of our systems.

The first modem I ever used was capable of something less than 300 bits per second. The first one I bought could manage 2400. The next, 9600.

The one I have at home now is limited only by the bandwidth of my phone line/exchange/ISP, and currently tops out at a pathetic one and a half million. Some people who live a mile or so away from me are regularly able to receive forty million. Speeds in technologically advanced countries like South Korea top out over a hundred million, and rising.

Processing speeds have advanced similarly, but so has what we're trying to do, quite often. We expect to exchange high definition video files and complex three dimensional models quickly, where once we were yawning waiting for a text file to transfer.

I used to date a complex three dimensional model.

But she dumped me when I discovered h2g2 and got
involved with shallow 2 dimensional entities of
unknown politics and questionable gender.

~jwf~