Slashdot Mirror

Does anyone know of a good source of fresh statistics for connection speeds? For general web stats, The Counter has decent free aggregate stats for things like browser, OS, monitor resolution, etc. But I'd really like to find something similar for bandwidth speeds. Any ideas?

I feel a need to clear some things up here:

You shouldn't be forcing people to use a perticular font on a web site in any case whatsoever.

I don't force anything. CSS font selections are at best a suggestion,and if the font I've suggested doesn't exist, the default renders anyway. But 99% of the audience never changes their default font, which is Times New Roman, and which is not always the right font for the job.

That's a decision to be left to the client's renderer.

If it means that much to you, most modern browsers have the option to let local settings override the designer's settings. Edit your preferences and enjoy! Most people don't choose their default font, and don't care if the font changes as long as the design is usable (and yes, I have the data to back this up from usability studies I've done).

If you absolutely MUST dictate a precise font face, for a specific purpose, then create a graphic of what you want, and use that.

And that's more friendly how? Plain HTML text is searchable, cut-and-pasteable, and accessible in web readers for the blind, etc. Text-as-graphics is a last resort that should only be used for logos and such.

What about sight-impaired people who need HUGE fonts in order to read a web page? If you've forced 12 pt blahfoobar, they're screwed.

Who said anything about forcing a font size? We're talking about font face, which you can specify while leaving default sizes intact. And when using CSS and the em unit, you can scale an entire page relative to the viewer's base font size. In fact this is another argument against text-in-graphics, as graphics always render as fixed pixel sizes, and limit the scalablity of designs.

And on the subject of font sizes, sometimes there's a need to make a super-small piece of text, as in legal copy or photo captions, things that are required to be on a page by law but will be read by few if any of the users.

As was noted in another post, Verdana was designed by Matthew Carter (the same typographer who designed Bell Gothic for legibility in the cramped confines of telephone books) specifically for legibility on screen. It's a welcome alternative to Times New Roman. Typography textbooks will tell you that serif fonts are more legible in body copy, but that assumes enough resolution to render the serifs! Current screens set to 800x600 don't compare to the resolution of print which can commonly run up to 3250dpi (that's a 1000x difference when comparing number of dots per page).

Having a font on demand (Verdana) with a minimum of visual crap around the edges is a valuable tool in the design arsenal.

Where do you get those browser-statistics? My source is TheCounter.com

Hmmm...

My own logs roughly jive with yours, but even more telling is TheCounter.com's global stats here

I always use thecounter.com's stats for making web dev decisions (should I not worry about writing for 640x480, or is that still a signifigant chunck of users out there?) since they represent roughly half a billion unique hits each month over thousands of web sites.

Two bits of trivia I do find funny:

- XP has fewer users than Linux or Web TV.
- Linux has fewer users than Windows 3.x!

I totally disagree with the Coward and his attitude (that's nice, calling someone a moron because you don't agree with him). Many clients say "I want these features and I don't give a #@$% about Netscape 4.08 since it's gonna cost me more to have you develop it. I'm happy to satisfy 9?% of my possible audience."

Others say "I want it to do everything (DHTML, CSS, ActiveX, Flash, integrated Authorization and Authentication, SSL etc.) with every browser" until we tell them the price of the development, and the potential bugginess....

It's easy enough to say "make it standards compliant", but the different browsers implement standards differently Take CSS, for example, and how about printing? Why do you think there are so many pages devoted to cross browser functionality? BECAUSE IT'S HARD AND TAKES TIME. TIME MEANS IT COSTS THE CLIENT.

Not every client has the $ resources of an Amazon or an Ebay. Do you work for real live clients?

TheCounter Browser Stats for March 2002. By the way, what is the best method to detect Mozilla on the UserAgent? Look for "Gecko"?

http://www.thecounter.com/stats/2002/January/os.ph p

More accurate data about Google visitors can be found at the august Google Zeitgeist. It's a few months out of date, but anyway, the number given there is 1.18% percent. Seems that Google has 150 million queries per day, which is nearly twice the amount of hitbox visitors. (of course I don't know if those two terms mean the same in this case...) I'd say that all in all, Google is at least as reliable source of statistics for this matter than Statmarket. (whether either number is very reliable is up to someone with real knowledge of math to discuss ;)

Just for fun, let's mention that TheCounter.com seems to give a value of about 0.27% for the current month, although again, I'd not trust the number very absolutely any more than Google or Statmarket.

http://www.thecounter.com/stats/2001/December/os.p hp shows Linux below even Windows 3.x and WebTV

I'd like to know what legal ground you would stand on if you claimed you didn't have a javascript-capable browser, or had disabled scripting (this isn't a rare thing - 12% of users according to thecounter), and therefore could not read their disclaimer page. A lot of websites use scripted popups to open such legal things (and more important ones too than this garbage).

It's not that Linux will die, it's just that it is going to remain insignificant on the desktop as long as it it fails to address its usability problems -- many of which have nothing to do with the perennial command line vs. GUI debate! Instead they have to do with things like device capability databases and backward compatibility testing.

(I'm not disagreeing with you here, just expanding a bit on your comments.)

Tim

--

are up to date, at around a million hits per hour.

IE 4 and NS 4 are both around 10%. This means they can't be ignored. But if they fall a few percentage points more marketing will forget abot them (this happened to VGA screen sizes).

TheCounter.com Browser statistics

The big problem is that Netscape 4 still has around 9% of the market - this needs to fall further before it can be ignored.

VGA resolution didn't get ignored until it was around 8%.

Last time I checked some statistics, the various variants of Windows ran 95% of the browsers, Mac 4% and "Others" one single percent. Thats Linux, Solaris, BSD etc combined!

More current data, using TheCounter.com statistics give roughly the following percentages:

• Win 98: 66.9%
• Win 95: 15.0%
• Win NT: 7.2%
• Win 2000: 3.9%
• Unknown: 3.7%
• Mac: 2.1%
• WebTV: 0.6%
• Linux: 0.3%
• Unix: 0.2%
• Win3.11: 0.2%
• OS/2: less than 0.1%
• Amiga: less than 0.1%

Just for a test, I calculated the same stuff from last years data:

• Windows 98: 55.5%
• Windows 95: 30.7%
• Windows NT: 8.5%
• Macintosh: 2.1%
• Unknown: 1.3%
• WebTV: 0.7%
• Windows 3.x: 0.6%
• Misc. Unix: 0.5%
• OS/2: less than 0.1%
• Amiga: less than 0.1%

(I'm trying to avoid facing the less than interesting studying project that I should start, that's why I bothered to calculate all those fancy numbers ;)

If someone is worried about lamers stealing their content, then there are several ways to combat it:

1. Strongly brand your site. Put big notices up telling people that your service is free and they're being ripped off if they're paying for it.
2. Use cookies. Issue a cookie from the home page. If people don't have the cookie redirect them through an "about our site" page before letting them go to the link.
3. Embed session ids into the URL. Lamers can't link to a URL if it changes all the time.
4. If the lamer is loading your content into a subframe, use Javascript to force your content to the whole window.

If you really want to see how referrers are an invasion of privacy, try putting a sophisticated web counter into your homepage and read the statistics that it gathers about your friends when they visit. Next time you're out on the town you'll be able to ask which of them was visiting www.xxxgaybondage.com before they went to your site.

Is anyone else annoyed by the big Internet.com "network"? It seems that they are going around and buying every single worthwhile and half-worthwhile website out there? After the site is acquired, it becomes slower than it was before, because they fill it with ads and links to their other sites.

During the last year, this has happened to many of the sites I frequent. I'm getting really bugged, and a even a little scared. This post on Slashdot just gives me another reason to be so.

• TheCounter was an excellent counter service. They had a website with two banner ads on each page, and a nice clean layout. The counters they provided loaded quickly and reliably. After having been acquired, the website was redone to something that makes my eyes hurt, and the counters started either taking ages to load, or failing to load at all. Their solution? Create a new account so that it exists on our new server.
• No major changes yet at LinuxApps, except for all the links and ads for the other Internet.com sites. However, the same problem as the one listed in this post arises. Would a post about Freshmeat on LinuxNewbie be pulled because Internet.com owns LinuxApps, which is similar to Freshmeat? This is, apparently, what happened in this case.
• BrowserWatch, LinuxNewbie, PHPBuilder, and a lot of other sites I frequent have all been bought by Internet.com. There haven't been major changes on them yet, besides all the annoying links and logos, but this post gives me more reason to be paranoid.

--