adrian holovaty

September 23, 2002, 9:24 PM ET

Emulate really old browsers

No site review tonight; I'm busy converting and backing up all my guitar/piano recordings from the past three years.

But in the meantime, check out dejavu.org's browser emulator. It lets you see what any Web page would look like in vintage browsers such as NCSA Mosaic and Mosaic Netscape 0.9 -- all while duplicating the original browsers' interfaces quite faithfully. Fascinating. (Thanks to Stuart for the link.)

Comments (4) / Permalink

September 23, 2002, 12:18 PM ET

Monday's recommended links

IBM has an article on XHTML 2.0.

Two weeks ago, the Web site of the Recording Industry Association of America was hacked a few times. Now we know how they did it. The Register reports it was discovered that the site's robots.txt file (a file that gives robots instructions on which pages to ignore) pointed to the site's administration page, which wasn't password-protected. The lesson here: Never put anything in your robots.txt file that you don't want public, as I've said before. (Link from dive into mark)

And the big news of the day is that the new-and-improved Google News is now accessible to all users. Previously, it had been available only to selected users. (I was one of the lucky ones and reported this last week.) Everybody's talking about it: usability experts, webmasters, journalists and even Holovaty.com readers. For more, see the Reuters story and News.com story. Also, the ResourceShelf has more on what exactly is new about the site.

In the News.com story, a Google product manager who developed the site is dead-on about why this is an important innovation:

[T]his changes news reading habits because (usually) you pick a source and pick the story that interests you. With this service, you pick the story that interests you and then pick the news source.

Comments (5) / Permalink

September 23, 2002, 12:09 AM ET

Holovaty.com site improvements

Today I've made a few changes under the hood of my content-management system, with the intention of making this site faster and easier to use:

• All pages are now automatically compressed using gzip. That means my server compresses the content and your browser decompresses it behind the scenes, leading to faster download times for you and less bandwidth use on my end. Most modern browsers support this type of compression; of course, the ones that don't can still access this site. (Evidently gzip can reduce download time by up to 50 percent, although I'm on a broadband connection and honestly can't tell, because it was fast already. Anyone out there feel a difference?)
• I've implemented what I call smart anchoring, which is basically an artificially intelligent means of linking blog entries to one another. It works like this: If a blog entry links to a previous entry that happens to be on the same page (e.g. if both entries are on the current home page), the link will be a simple in-page link that jumps you down to the linked entry automatically, instead of reloading the entry in a new page. Otherwise, the link will take you to a new page. For example, this link points to the previous blog entry. If you're viewing this entry from my home page or the September 2002 archive page and both entries are on that page together, it'll jump you down the page. Otherwise, it'll take you to the permalinked version of that page. I've never seen this feature on any other site, so I'm quite pleased with it.
• You now have the option to toggle search-term highlighting. I've had search-term highlighting in place here for a while, but I've added a checkbox on the search results page that lets you toggle it. It uses client-side scripting, so it won't appear if your browser doesn't support the JavaScript functionality it uses. (I'm a firm believer in showing users only what their browser can support; it's annoying to click something only to be insulted with a "Your browser doesn't support this" pop-up.)
• Finally, the search engine accepts quoted strings. So now you can search for "user registration" instead of just user registration. (The former would find only blog entries with the words "user registration" next to each other in that order, while the latter would find any blog entry with both those words, regardless of order.)

Comments (3) / Permalink