Recently I was re-watching Cringely's: "Triumph of the Nerds: Great Artists Steal" --part three of a documentary about the rise of the personal computer. Something Bob said bugged me, and it's been nagging me since. Here's the quote (referring to the difference between GUI and command line interfaces):

"Pictures rather than words, making the pc easy and intuitive."

It seems to be commonly accepted that the reason graphical user interfaces have been successful is because they are more "intuitive" (whatever that means.) The truth is, it's not that graphics are more intuitive than words, it's that the gui by it's nature has a more limited command set, thus making it easier to learn...

But don't take my word for it, here's what user interface guru Alan Cooper says in his book "About Face: The Essentials Of User Interface Design": "The first commercially successful implementation of the PARC GUI was the Apple Macintosh, with it's desktop, wastebasket, overlapping sheets of paper and file folders. The Mac didn't succeed because of these metaphors, however, but because it was the first computer that defined a tightly restricted vocabulary--a canonical vocabulary based on a very small set of mouse actions--for communicating with users. It also offered richer visual interaction. The metaphors were just nice paintings on the walls of a well-designed house."

Cooper is specifically referring to the Mac, but of course the same also applies to Microsoft's less elegant Windows GUI.

Don't get me wrong, Cringely is a cool guy, and what he said is true enough in the context of his documentary --just not completely accurate.

Most pictures are not that great on web pages. I mean, everybody understands that they are purposely degraded so that they are "modem friendly". So when I see a picture that looks interesting and I can't quite make out the detail I will run the cursor over the pic to see if it's linked. What do most people expect to find when they click a picture that is a link? What I expect to find is a bigger less "modem friendly" version of the picture! What I usually get though is a link to the story. If I wanted that I would click on the text link to the story!

I mean... can there be any more fundamental violation of design? Why then is it so commonly done? Most of the big newpaper sites are guilty of this.

Here's an example of what I mean:

http://www.apple.com/macosx/

See that screenshot of Mac OS X in action? Kind of a tease right? You can barely make out what's going on in the shot. Wouldn't you expect clicking on the picture to take you to something where you could see what was going on? Of course you would! What do you get instead? More text and shitty graphics!

Like I heard some vegetarian saying about hotdogs... just because everybody does it doesn't mean it's not gross.

I like the quote: " What that means, basically, is that it's hard to speak and think at the same time." I know just what you mean... Interesting article. Could have gone into a bit more detail. Washington Post: A Visual Rather Than Verbal FutureEarlier I was kind of half watching Charlie Rose. He interviewed a couple of executives that "suffered" from dyslexia and a doctor. One of the things that kept coming up was that there are many successful people that are dyslexic, and yet the doctor kept talking about it being a disfunction. It struck me [again] how arbitrary the definition of normal is. The doctor's perspective was that those people were exceptional BECAUSE they were over compensating for a deficit they had.

Looking at Lubrich's Arts Journal and noticing that there seemed to be a lot of pictures on his frontpage. Wondering about load time. Then I was looking at that site I did for Dave and noticed the bookie picture of him is 60k. Way too big for a black and white image. Anyway all this reminded me that some people look at my site using a dialup connection. Was wondering how total page size affects load time.

Here's a comparison of famous weblogs:

Site:	Size:	Load Time:
preterosso.mine.nu	18k	4 sec
preterosso.dyndns.org	20k	4 sec
rudedog.mine.nu	51k	10 sec
knurdle.com	53k	11 sec
lizardspeaks.ath.cx	58k	12 sec
subminimal.org	114k	17 sec
kacked.com	86k	23 sec
lubrichs.mine.nu	256k	51 sec
(dl times based on 5k/second optimum speed of a 56k modem)

Usability guru Jacob Neilsen writes: "...the minimum goal for response times should therefore be to get pages to users in no more than ten seconds, since that's the limit of people's ability to keep their attention focused while waiting..."

Alertbox: The need for download speed

michael|64.236.243.243|preterosso@adelphia.net|http://preterosso.mine.nu|2|3|19|2002|3|27|23|PM|The Horror, the horror... sunil|128.125.227.62|||2|3|19|2002|6|58|54|PM|The data is intrinsically flawed. The results should be depicted in a statistical format. Mean and Standard deviation information should be included. Basically the load times should be checked throughout the day and over a period of a few days, and then the results should be averaged out sunil|128.125.227.62|||2|3|19|2002|7|3|39|PM|In addition to my previous post, the reason for my suggestion for statistical measures is that internet traffic is intrinsically bursty in nature. Please look at one example of the nature of internet traffic in the following paper published in the IEEE Trans. on ACM:

Mark E. Crovella and Azer Bestavros, "Self-Similarity in World Wide Web Traffic: Evidence and Possible Causes", in IEEE/ACM Transactions on Networking, 5(6):835--846, December 1997 John|192.168.0.2||http://kacked.com|2|3|19|2002|7|13|58|PM|That might be true (it might also not be relevant... heh!) My intention was to stimulate discussion and try to promote understanding of the experience of the dialup user. Certainly for today the results are valid. Tomorrow (if I post and that 16k blockheads graphic scrolls into the archives) the results will be different --but not by that much. You're welcome to take the time for a complete statistical analysis, however, I'm more interested in your day to day experience and impression of load time as a user of these sites. That's where the real data is, since other factors besides page size must be considered, such as: server response time, server downtime, etc. Since there is no easy way of getting or approximating this information the best thing is to garner the impressions of the users. What I'm trying for after all is a favorable impression with my readers, so why not just ask? I guess I wasn't being clear but that's what I was after.

So. That being said, what's your impression of the relative speed at which these pages load? I'm assuming that sometimes you browse these sites using a 56k modem... sunil|128.125.227.62|||2|3|19|2002|7|29|40|PM|The data that you supplied otherwise is quite interesting. I tried loading yours several times and I am getting 10-11 sec now! Quite a difference, and I am using 56K modem. But definitely this stuff you posted is really quite interesting! sunil|65.208.137.206|||4|3|21|2002|3|22|7|PM|BTW I am still wondering as to why my computed numbers (optimal and non-statistical) don't match those numbers that you posted (so it is still meaningful)...see my post at the next topic