[Bob Williams]

Originally Posted by Forrest

I think that sums up a clear difference in our opinions. I think the text version is just reporting the facts, and that any sort of bar that acts as a measure of those facts, is just a guess. I would rather the guesswork be left up to the user.

I'm still missing you here, though. If the text version says "5 of 10", you don't really care that 5 resources have loaded, and you don't really care that there are 10 resources in total (at the moment) - those are transient facts that, in themselves, aren't usually important. What you're interested in is the relationship between them, and that's why they're presented together--with a word ('of') that strongly denotes a relationship, no less--without even the option to do otherwise. If you're familiar with basic statistics, it's like looking at two sets of results versus looking at the correlation between them; each set of results on its own is of limited usefulness, while determining the relationship between them is very useful.

Further, any conclusion drawn on the relationship between x and y is mathematically just as much fact as x and y themselves. You can even swap things around them, showing something like "5 resources loaded (50%)", which is enough to derive that y is 10. We're talking about equivalent things in a literal, mathematical sense. Massaged algebraically, yes, but still equivalent. A progress bar doesn't change this at all; it simple gives the information visually. You may not know the exact numbers (though some progress bars do provide them), but the human brian can process that visual feedback in a very similar way, and when the numbers themselves aren't important, the visuals are good enough. In fact, progress bars show the relationship better (in a human interpretive speed sense) than the two numbers alone - which is why OSes are filled with progress bars instead of simple pairs of numbers.

In short, the only guesswork is about the true total resource count, and that remains just as much a guess whether you're using a textual or visual representation. Beyond that single guess, you're solidly into the realm of guess-free mathematics. And since OW doesn't try to make that guess, there's simply no guesswork going on at all.

That's really weird. I do what you often do as well, but never have it launch in list view then switch. I don't use list view, and the only reason I can think of using it is to conserve space.

I very rarely use list view of my own accord (it's more efficient for lots of tabs, but I just like the nice-looking thumbnails too much to give them up), but OW regularly switches to them on me at startup. Well, more accurately, it just starts up with them. Not always, but often. As for the switch back, sometimes it happens when I start switching among the tabs, and sometimes it just happens on its own. Sometimes it even does them as I use them. It is indeed a bit weird. I figured it was just OG's way of conserving processor time by not trying to render several dozens of tabs simultaneously, but perhaps not.

Would adding a progress bar to those tabs be useful for the people who are concerned about space?

I don't see why not. It need not be big. In fact, it need not even be the standard scroll thumb-like bubble blob (hey, if Safari can get away with it...). I can visualize a 2px or so high bar being included right under the name; this probably wouldn't require making the tabs any taller at all, or if so, then by just a pixel or two, which I don't think many folks would argue against.

I disagree on that. Obviously it has a great deal to do with the sites a user visits, but my bet would be that a great number - possibly the majority - of sites don't have all the resources linked to from within the HTML. Very often site have a linked CSS file which also links to images.

I didn't say all resources tend to be linked in the HTML file. I said the page's body text is usually in the HTML, and I stand by that. On a well-designed page, the primary purpose of the HTML is to serve as a vessel for the text content. On poorly designed pages (most, unfortunately), the HTML still almost always serves this purpose, but it also serves to provide much of the formatting information.

I have seen software all to often try and be smart about things, and end up looking dumb in the process. Look at Windows and it's "unknown" errors. They're trying to be smart about telling the user what the error was, but all too often the error is "unknown."

I agree we don't want OW jumping to conclusions on us, especially if those conclusions will often be incorrect. But a progress bar representing the x of y resource status is does no such thing; it simply presents the same information in a different way. In either case, it's up to the user to not jump to conclusions and read the data for what it actually is.

Safari's implementation is a good example of another re-hash of using a progress bar to denote how far a page has loaded. It can start off with a quarter inch, within a second shoot to 90% full then sit there for another ten seconds.

Yup, and OW's x of y may jump to 9 of 10 then sit there for ten seconds. What's the difference? Assuming you don't care that the browser believes there to be 10 resources at the moment and that it's loaded 9 of them, they're the same thing. Again, as long as you don't try to interpret either one directly as a measure of time, then they both are useful. Admittedly, it's easier for a naive user to fall into the trap of assuming a progress bar relates to time than an x of y, but not by a whole lot. And when most free, OS-provided browsers do it, it's a hard argument to make that OW should protect naive users :-).