Latest Computer Comparison

At 7:21 AM -0700 2007/06/29, Edward Reid wrote:

>Looking at my current memory use (XP, but I don't expect that OSX is that
>much different), I see that the largest user is one you didn't mention --
>the web browser (Opera), with 150MB. Next comes ... Eudora! with half that.
>And Adobe Reader is close on Eudora's heels. Now at least I can understand
>a web browser using a lot of memory -- they are one of the most intense CPU
>users today and provide real benefits, and using more RAM to speed them up
>is worthwhile. But 77MB for Eudora? I don't generally even display images
>in email! And some parts of Adobe Reader slowed down dramatically on v6 and
>have not come close to recovering -- AFAICT the slow-down has brought me
>absolutely nothing except lost time.

        AIUI, for every message you have open onscreen, Eudora keeps
the whole mailbox (or at least its TOC) in memory. Many people keep
lots of unused messages open for later, which are only read in at
launch and then paged out, never to be touched again until next
launch. The ones which are used are loaded into memory -- it's called
"caching". ;)


                                                Chris
--
Chris Pepper: <http://www.reppep.com/~pepper/>
SSL for Surfers & Sys Admins: <http://www.reppep.com/~pepper/ssl/>

On 6/30/07 2:21 AM, "Edward Reid" <edwardpaleo.org> spake thus:

> Response Time and Display Rate in Human Performance with Computers, by Ben
> Scheiderman, Computing Surveys, Vol. 16, No. 3, September 1984, pp 265-285.

Aha! Yes, this looks like the one:

MILLER, R. B. 1968. Response time in man-computer conversational
transactions. In Proceedings of the Spring Joint Computer Conference, vol.
33, pp 267-277.

According to Shneiderman and a couple of other sources I've picked up,
Miller discusses three response time thresholds based on the way the brain
woks: 0.1s or less is considered instantaneous, 1s or less is considered
"interactive" and 10s or less is sufficient to maintain focus on the task
(which is probably why I find it harder to multitask these days --- things
are finishing in less than ten seconds). Miller recommends a general
response time of 2s.

Unfortunately not available electronically, and my library doesn't have it
either :( Then again, Shneiderman is probably authoritative enough for
most people.

--
Nigel Stanger, Dunedin, NEW ZEALAND.
http://xri.net/=nigel.stanger

John Welch wrote:
> To even begin to believe the "vast majority" of computer users live a
> completely single-task life, I'm going to need a freight train of proof.

I have no idea about the "vast majority". But I do know it's a non trivial number. I consult in small offices. I'm there a lot and get to watch people work. You can tell the single task folks easily. They maximize EVERY window except maybe Finder windows. They fill their 20" iMac display with a single email. If they open an attachment they have the result fill the display.

Others like me open up 20 overlapping windows and switch between them and applications non stop.

And some folks operate in the middle of this spectrum.

What I strongly believe, and most folks tell me I'm wrong, is that this is very much related to age. Not totally the age of the person but the age at which they started to use a computer and what they did with it. It is closely related to the age of the person but not totally. Most 20 somethings are fierce multi-taskers. Most people who learned to type on a real typewriter are steadfast single taskers. People who got well into their professional life before they got into computers tend to be single taskers.

At least that what I've observed over the last 10 years or so that computer have allowed easy multi-window multi-tasking.

David Ross

On 03/07/2007 8:40 AM, "John Welch" <jwelchbynkii.com> spake thus:

> To even begin to believe the "vast majority" of computer users live a
> completely single-task life, I'm going to need a freight train of proof.

There's a distinction here between multitasking by the computer (process
scheduling) and the user (workflow). *Of course* the computer multitasks ---
no-one's disputing that. But whether the *person* using it multitasks (i.e.,
does the person work in a manner that takes advantage of the computer's
multitasking) --- that's not so clear. Just because I leave my email client
open in the background doesn't imply that I'm paying any attention to it.
That strikes me more as a form of delegation than multitasking per se.

--
Nigel Stanger, Dunedin, NEW ZEALAND.
http://xri.net/=nigel.stanger

Nigel Stanger wrote:
> On 6/30/07 2:21 AM, "Edward Reid" <edwardpaleo.org> spake thus:
>
>> Response Time and Display Rate in Human Performance with Computers, by Ben
>> Scheiderman, Computing Surveys, Vol. 16, No. 3, September 1984, pp 265-285.
>
> Aha! Yes, this looks like the one:
>
> MILLER, R. B. 1968. Response time in man-computer conversational
> transactions. In Proceedings of the Spring Joint Computer Conference, vol.
> 33, pp 267-277.
>
> According to Shneiderman and a couple of other sources I've picked up,
> Miller discusses three response time thresholds based on the way the brain
> woks: 0.1s or less is considered instantaneous, 1s or less is considered
> "interactive" and 10s or less is sufficient to maintain focus on the task
> (which is probably why I find it harder to multitask these days --- things
> are finishing in less than ten seconds). Miller recommends a general
> response time of 2s.

I distinctly remember the discussion in 1980 where we were working with the 1 and 3 second guidelines in our software. This was a very small company compared to many software companies these days. There was the founder who wrote much of the original system and me who was doing about 60% of the development at the time.

So either the above paper was proof of the general knowledge of someone did something earlier.

David Ross

on 7/2/07 4:40 pm, John Welch at jwelchbynkii.com wrote:

> I'd argue that the vast majority of people do multitask, whether they want
> to call it that or not. At the very least, email and browsing, with a
> slightly smaller percentage having an IM window open.
>
> I highly, HIGHLY doubt, and will until I see some proof with a seriously
> large sample size, that the vast majority of people open their email, do
> some email, then shut it down. Then they open a web browser, and ONLY DO
> THAT, then shut it down. Then re-open email. If there's a link in the email
> they click on, as soon as the browser opens, they then shut down their email
> before proceeding. If they need to do a little non-email letter writing,
> they then shut down EVERY OTHER PROGRAM before opening that word processor.

What you're describing is the extreme end of single-tasking, reductio ad
absurdium. But not doing that is not necessarily "multitasking". If I read
an email, click on a link in it, wait for that link to load, view that page,
then go back to reading email, I haven't really "multitasked" just because I
didn't quit the email program in the middle of that sequence. Viewing the
web page was a subtask of reading the email that contained the link.

I tend to leave programs open after I've launched them, whether I'm using
them or not, until I log out. Does that mean I'm actively multitasking that
whole time? I seriously doubt it. For me, "multitasking" means I've got more
than one simultaneous activity that I'm paying active attention to, either
concurrently or in rotation (and I normally don't count listening to music
as part of that--for me, that's background noise). Moving from one program
to another without closing the first doesn't automatically fill that bill.

Mark D. McKean
qpandaquantumpanda.com

On 2-Jul-2007, at 14:40, John Welch wrote:
> On 6/29/07 09:21 AM, "Google Kreme" <gkremegmail.com> wrote:
>
>>> However, that's assuming that all you do is sit there and watch
>>> your browser
>>> work. Multitasking means you hit the site, then go do something
>>> else while
>>> it loads, which is something the original comparison didn't even
>>> mention.
>>> That's just one reason why it's such a load of silly.
>>
>> The vast majority of people do not work this way, however.
>
> I'd argue that the vast majority of people do multitask, whether
> they want
> to call it that or not. At the very least, email and browsing, with a
> slightly smaller percentage having an IM window open.

Having three apps open is not necessarily multi tasking. It's more
like task switching. I watch people quite a bit, and for example,
someone clicks on their 'check mail' button and then they wait while
the client checks their email. If it only takes 20 seconds, or even
a minute, they don't do something else, they just wait for that task
to finish.

When they are loading a webpage (say a movie trailer) the same
thing. A 720p HD trailer, even on a good wifi connection, may easily
take 60 seconds to start playing, and most people sit and wait for it.

They switch from one completed task to another completed task and it
is the rare geek that will flip around from task to task to task
leaving one half finished to go to something else half finished and
onward.

> I highly, HIGHLY doubt, and will until I see some proof with a
> seriously
> large sample size, that the vast majority of people open their
> email, do
> some email, then shut it down. Then they open a web browser, and
> ONLY DO
> THAT, then shut it down. Then re-open email. If there's a link in
> the email
> they click on, as soon as the browser opens, they then shut down
> their email
> before proceeding.

They don't shut it down, but they complete the task they are working
on before going on to something else.

Read email, click link, watch funny video, close browser window (only
the really advanced will simply switch back tot he email), read next
email, reply, read next 4 emails, click a link, look at funny
cartoon, close window, go back to email, read the last 3 messages.
OK, done with email. Now, go the web browser and load some pages.

As opposed to the multitasker whose normal process is more like:

read email, click link, watch funny video, switch back to email,
start replying to an email, switch to browser, look something up. Go
back to email, read some more, add something to the email you're
replying to. Load a couple other web pages. Load up your editor of
choice to make some changes to a webpage. Start downloading an update
to CoolApp.app and the newest BigMovie trailer, go back to email,
finish replying to that message and send it off. Check for more
messages. Install updated CoolApp.app and watch the movie trailer you
finished downloading. See if any new mail came in.

The latter is multi-tasking, the former is task-switching. Wait
until the current task is done and then switch to another task, leave
very few task incomplete, and always go right back to the most
recently incomplete task. The multitasker will leave many tasks in
process at any one time, and may have unfinished emails, unread
webpages, and a variety of open documents all of which are current
tasks.

About the only really multi-tasking most users will do is IM, where
they will bounce between IM and whatever else they are doing.

NB: I'm not saying one is better than the other. I am a multitasker,
but another word for that is 'scatterbrain' and it takes me longer,
sometimes MUCH longer, to complete any one task. for example, whilst
writing this email I've been running some rysnc backups and grepping
through a month of logs looking for something. If I did the grep
pattern right, I have the information in a terminal window just
waiting for me to switch to it. If not, I've wasted a good 15
minutes with the wrong results sitting there waiting for me to look
at them.

> If they need to do a little non-email letter writing,
> they then shut down EVERY OTHER PROGRAM before opening that word
> processor.

Shutting down the apps has nothing to do with it.

On Jul 2, 2007, at 1:40 PM, John Welch wrote:

> I'd argue that the vast majority of people do multitask, whether
> they want to call it that or not. At the very least, email and browsing, with a
> slightly smaller percentage having an IM window open.
>
> I highly, HIGHLY doubt, and will until I see some proof with a seriously
> large sample size, that the vast majority of people open their email, do
> some email, then shut it down. Then they open a web browser, and ONLY DO
> THAT, then shut it down. Then re-open email. If there's a link in the email
> they click on, as soon as the browser opens, they then shut down their email
> before proceeding. If they need to do a little non-email letter writing,
> they then shut down EVERY OTHER PROGRAM before opening that word
> processor.

The above mostly doesn't address whether the USER is multi-tasking.
It addresses whether the computer is.

I seldom read email while *actively* conversing in IM. I seldom pay
attention to both the email program and the web browser interleaved
over a short period (unless I'm looking something up in the browser
related to the current email message, or vice versa). I seldom have
more than one email message open at once (and perhaps 90% of the
situations in which I do, it's to check Received: headers to
determine where a seeming duplicate message happened--so it's only
one task).

So, primarily, *I* don't multitask. Mostly, I suspend one task
before turning to another. This computer does...it currently has
several tabs open in Safari, Yojimbo open (but being ignored--I was
just looking up a password), Little Secrets open (but being ignored,
as with Yojimbo), Terminal open (because I was experimenting with
something a couple of hours ago, iChat open (with nothing going on),
Mail.app open (working on this), Preview open (but with no windows
open--left over from earlier), Adobe Reader open (I was browsing the
iPhone User Guide an hour ago while preparing a page in our support
Wiki).

   --John

On Jul 2, 2007, at 1:48 PM, Nigel Stanger wrote:

> MILLER, R. B. 1968. Response time in man-computer conversational
> transactions. In Proceedings of the Spring Joint Computer
> Conference, vol. 33, pp 267-277.
>
>
> Unfortunately not available electronically, and my library doesn't
> have it either :( Then again, Shneiderman is probably authoritative
> enough for most people.

I'm not sure that report matters any more. The "man" (person) in
"man-computer conversational transactions" these days is "nearly
everyone". Then, it was a tiny minority of everyone, and they were
being paid to do the job of conversing with the computer. (I believe
I am right in remembering that we just passed the 40th anniversary of
the ATM, which was one of the first spreadings of the conversing
population. And ATMs were the exception for a long time.)

I can, however, give an example of "too slow". At NCR in the late
1970s, we did our software development work on a CDC Cyber 175.
Although the machine was a node on Cybernet, it was exclusive to NCR--
it was a node so NCR world wide could use it. It was also the
busiest node on Cybernet, and taught CDC a lot about tuning a Cyber
175. At one point, using a line oriented editor took a 5 minute wait
after hitting return before the result came back. (That day, I went
home and downloaded the file I was working on onto my Apple ][ and
used the UCSD Pascal editor to edit it, then uploaded it.

Except that I told the CDC guys--whom I knew pretty well--that I had
"uploaded" to my Apple and later "downloaded" to the Cyber. When
they objected, claiming I had that backwards, I pointed out that
"uploading" is sending the file to the more useful machine.

The did get the node tuned to give decently fast responses again, but
it took a while.

   --John

On 7/3/07 08:40 AM, "David Ross" <drdavidrossconsultant.com> wrote:

> I have no idea about the "vast majority". But I do know it's a non trivial
> number. I consult in small offices. I'm there a lot and get to watch people
> work. You can tell the single task folks easily. They maximize EVERY window
> except maybe Finder windows. They fill their 20" iMac display with a single
> email. If they open an attachment they have the result fill the display.

That's not single tasking, that's focusing on the task at hand. But leaving
your email running while you are on a web site or using Word? Still
multitasking. Multitasking does not require you to jump from window to
window like an over-caffeinated frog.

As far as maximizing windows, it's not only not uncommong, it's actually
standard SOP with Windows. In fact, it's SO common that the Office 2007 team
designed the Ribbon with that assumption in mind, because it's how Windows
users make the per-window menubar.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 7/3/07 08:40 AM, "Nigel Stanger" <nstangerinfoscience.otago.ac.nz>
wrote:

>> To even begin to believe the "vast majority" of computer users live a
>> completely single-task life, I'm going to need a freight train of proof.
>
> There's a distinction here between multitasking by the computer (process
> scheduling) and the user (workflow). *Of course* the computer multitasks ---
> no-one's disputing that. But whether the *person* using it multitasks (i.e.,
> does the person work in a manner that takes advantage of the computer's
> multitasking) --- that's not so clear. Just because I leave my email client
> open in the background doesn't imply that I'm paying any attention to it.
> That strikes me more as a form of delegation than multitasking per se.

It's still human multitasking, which is why the whole OMG UR W3B PAG3 TOOK 2
LONG time thing is so much less important. It means that you know you *can*,
with ease, go do something else whilst the page loads. If you choose to sit
there and do nothing but stare at a window until that one task is done,
that's an option, but in the nigh-20 years I've been doing support, I've not
seen that particular behavior in quite a long time, pretty much since
pre-emptive multasking became the default, and not something for the
computer elite.

I can also tell you that I am *not* talking about the most technically
competent people on the planet. Unless you think that having FOUR toolbars
in IE on a 14" monitor is a sign of l33t.
--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 7/3/07 08:40 AM, "Mark D. McKean" <qpandaquantumpanda.com> wrote:

>> I'd argue that the vast majority of people do multitask, whether they want
>> to call it that or not. At the very least, email and browsing, with a
>> slightly smaller percentage having an IM window open.
>>
>> I highly, HIGHLY doubt, and will until I see some proof with a seriously
>> large sample size, that the vast majority of people open their email, do
>> some email, then shut it down. Then they open a web browser, and ONLY DO
>> THAT, then shut it down. Then re-open email. If there's a link in the email
>> they click on, as soon as the browser opens, they then shut down their email
>> before proceeding. If they need to do a little non-email letter writing,
>> they then shut down EVERY OTHER PROGRAM before opening that word processor.
>
> What you're describing is the extreme end of single-tasking, reductio ad
> absurdium. But not doing that is not necessarily "multitasking". If I read
> an email, click on a link in it, wait for that link to load, view that page,
> then go back to reading email, I haven't really "multitasked" just because I
> didn't quit the email program in the middle of that sequence. Viewing the
> web page was a subtask of reading the email that contained the link.
>
> I tend to leave programs open after I've launched them, whether I'm using
> them or not, until I log out. Does that mean I'm actively multitasking that
> whole time? I seriously doubt it. For me, "multitasking" means I've got more
> than one simultaneous activity that I'm paying active attention to, either
> concurrently or in rotation (and I normally don't count listening to music
> as part of that--for me, that's background noise). Moving from one program
> to another without closing the first doesn't automatically fill that bill.

So now to be "multitasking" you essentially have to be using your computer
like a poster child for ADD? Single tasking on a computer is fairly simple
to define: you have one user-initiated application open at a time. (the OS's
background stuff doesn't count here.) You're doing one thing at a time. If
you have multiple user-initiated applications open, you're multitasking.
Doesn't matter that you're only *active* in one thing at a time. Most people
on the planet are not going to be truly simultaneously looking at two
separate windows/applications and/or typing in two separate
windows/applications. But when you've got your email running in the
background, and you're in the middle of doing something else? Multi-tasking.
Really.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 7/3/07 08:40 AM, "Google Kreme" <gkremegmail.com> wrote:

> The latter is multi-tasking, the former is task-switching. Wait
> until the current task is done and then switch to another task, leave
> very few task incomplete, and always go right back to the most
> recently incomplete task. The multitasker will leave many tasks in
> process at any one time, and may have unfinished emails, unread
> webpages, and a variety of open documents all of which are current
> tasks.

If you leave multiple user-initiated programs open and running, you're
multitasking. You can call the human work patterns whatever you like, and
try to say that unless you're pinging around like an insane cricket, you're
not using a modern computer any differently than a 128K Mac, but multiple
user - initiated programs open = multitasking.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On Jul 3, 2007, at 6:40 AM, David Ross wrote:

> What I strongly believe, and most folks tell me I'm wrong, is that
> this is very much related to age. Not totally the age of the person
> but the age at which they started to use a computer and what they
> did with it.

Well, there were two ways I used "the computer" in my early days.
One was sequence of single-task/multi-task/single task: prepare the
card deck, submit it, do something else (usually until morning), pick
up and analyze the results.

The other was hands on--which was definitely single tasking. (And
yes, I did use IBM 704 and 709 hands-on--loved the powered rocker
switches on the console (there was a single "reset" that set all 36
switches for a word to 0, by slowly moving the ones set to 1 back to
0)).

So I think you're probably correct.

   --John

On 7/3/07 08:53 AM, "johnbaxterlistsmac.com" <johnbaxterlistsmac.com>
wrote:

> So, primarily, *I* don't multitask. Mostly, I suspend one task
> before turning to another. This computer does...it currently has
> several tabs open in Safari, Yojimbo open (but being ignored--I was
> just looking up a password), Little Secrets open (but being ignored,
> as with Yojimbo), Terminal open (because I was experimenting with
> something a couple of hours ago, iChat open (with nothing going on),
> Mail.app open (working on this), Preview open (but with no windows
> open--left over from earlier), Adobe Reader open (I was browsing the
> iPhone User Guide an hour ago while preparing a page in our support
> Wiki).

Humans aren't innately multitasking creatures once you get past the
unconcious stuff...ie walking and talking. We task - switch like fiends, but
can you simultaneiously read two documents, perfectly hear two conversations
and write two separate letters with both hands? Not without TONS of
practice.

However, we do multitask with computers, at least as close as we can come to
it, because they make it easier to start a process that will take a bit and
not require intervention, and while that's going on, do something else.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 04/07/2007 3:07 AM, "John Welch" <jwelchbynkii.com> spake thus:

> Unless you think that having FOUR toolbars
> in IE on a 14" monitor is a sign of l33t.

Oh man, those guys r teh r0xx0rz, I gotta get me some of that.

;)

--
Nigel Stanger, Dunedin, NEW ZEALAND.
http://xri.net/=nigel.stanger

On 04/07/2007 1:40 AM, "David Ross" <drdavidrossconsultant.com> spake thus:


>> MILLER, R. B. 1968.
> I distinctly remember the discussion in 1980 [...]
> So either the above paper was proof of the general knowledge of someone did
> something earlier.

?? Confused. Did you perhaps read that as 1986 instead of 1968? That's the
only way I can parse that sentence.

--
Nigel Stanger, Dunedin, NEW ZEALAND.
http://xri.net/=nigel.stanger

On 04/07/2007 1:53 AM, "johnbaxterlistsmac.com" <johnbaxterlistsmac.com>
spake thus:

> I'm not sure that report matters any more.

Given that the results were (supposed to be) a function of how the brain
works, I don't see that changing radically in only 40 years. Presumably
computers were used for the experiments because of the flexibility they
provided compared to anything else (i.e., easy to set up and control the
environment exactly how you want it). It could just as well have been
"man-automobile interactions" if I understand the context of the research
correctly. I'll know more once I get a copy.

I always like going back to the original sources for things, as they usually
throw up a few surprises. For example, a lot of database professionals who
think they know how relational databases work would probably get some big
surprises from reading Codd's original paper. I know I did :)

--
Nigel Stanger, Dunedin, NEW ZEALAND.
http://xri.net/=nigel.stanger

Interesting thread, especially since I'm the guy who wrote the UCSD Pascal Editor. Some context might be interesting: UCSD Pascal was ported to a minicomputer from a Burroughs mainframe. We were used to fairly expensive computer time, and fairly rudimentary text editors. The reasons why UCSD Pascal was fast and responsive:

1. It was designed as a tool for the introductory programming courses, and a bunch of research showed that interactive computing helped students learn faster.

2. As soon as we self-hosted, the developers of the system were its most ardent critics. We spent our undergraduate years beating that system so that it ran on 64KB machines. There was no space for bloat. The goal was that no activity (other than compilation) took any "noticeable" amount of time. We were aware of the Miller paper back then, and as soon as we self-hosted we thought his numbers were about right. We really noticed the difference moving from a 19,200 baud terminal (=2,000 characters/second, or about 300ms to paint a normal text screen) to a memory-mapped screen (that repainted in under 20ms).

3. The editor was designed to help folks writing both Pascal programs (e.g. the system itself, student programs) and school papers (ahem, I was an undergrad). The 64KB limit was a very useful thing; it certainly helped programmers make tradeoffs between space and features.

Feature bloat is a natural thing, and I'm as guilty of it as the next hack. The original PowerPoint (from Forethought?) was a very streamlined app on the Mac. I can't name too many features in the current PowerPoint that really help make presentations "better" (i.e. help present information in a useful way), but lots of features that make presentations "prettier."

And on the topic at hand: from the perspective of someone who had to use single-tasking systems for years, the ability to run multiple applications in parallel was a wonderful addition. The first multi-tasking implementation on the Mac was an absolutely useful thing. And Borland's Sidekick product was popular because it let you "multi-task" some rudimentary TSR (terminate and stay resident) applets.

Because "Moore's Law" is becoming "Moore's Cores" (i.e. single thread performance will remain more or less static, but multi-threaded performance will scale over time), desktop applications either need to know how to use multiple threads, or multi-tasking (in the user sense) is the only way perceived performance will increase. Other than the low-hanging fruit (e.g. image processing software like DxO), I fear this is going to be a big problem.

For those that like computer history:

<http://www.jacobsschool.ucsd.edu/Pascal/>

Cheers,

Richard