The front page of Slashdot on April 1^st, 2006

FTER Slashdot’s transition to proper CSS-based layout (it used to be heavily based on tables), a contest is run to select the new, permanent site design. Here are some of the latest design contenders. Some of these designs are included as actual Web pages, whereas others are appended as PNG-formatted screenshots. Shown above is the Slashdot front page as of April 1^st this year.

AVE you ever thought of an (electronic) article that you once saw or read? Could you find it easily after a few weeks, months, or years? How about if you bookmarked it? Would it always allow the article to be found again? And if so, how quickly can it be found?

Mozilla Firefox offers some good facilities for previously-browsed pages to be found. Both History and Bookmarks in Firefox have a ‘find as you type’ widget. However, this only contains and uses page titles.

What would make a better feature that is able to perform a more fuzzy and full-body search? The first option is to permit a search engine to spy on you and retain a record of pages that you follow. In turn, it will index these in isolation and allow your pages of choice to be searched cohesively. It is as though the knowledge conveyed in the index partly corresponds to what you already know. So, it is somewhat of a brain search, which makes the assumption that you carefully read every Web page you visited.

The option which exposes the user to a lesser privacy invasions is this: allow the Web browser to index a page whenever one is visited or only once at the end of the day. The index is of course cumulative and it permits the user to search just a ‘subnet’ — that which he/she has already explored.

Links can lose their value and
get rusty, even with Google

was innocently browsing the Internet this morning. By serendipity, I then arrived at a page where Google boast their contributions to Open Source software through funding (Summer of Code). It is only one example among others. But then, upon immediate inspection, merely all links turned out to be rel="nofollow"ed. SearchStatus made it evident by highlighting those links with red shades.

I have always adamantly believed that the purpose of this new class for links was different. I thought it was introduced in order to prevent and deter spam, among other things such as accommodation for microsformats (e.g. XFN). Here is the snag: If Google themselves are using rel="nofollow" to obstruct dynamicity into relevant, on-topic links, why should anyone else be hesitant to do so? rel="nofollow", a concept that was put in place by Google, is confirmed to have become something that can be misused. Its use has gone beyond the so-called ‘link condom’ (for spam) utility. Otherwise, Google demonstrate hypocrisy herein.

Yahoo search results for “roy” with
thumbnails (previews) added ‘on the fly’

AST year (July 2005 to be precise), I conceived a hypermedia system/Web browser wherein page previews get put adjacent to incoming/outgoing links. Such previews should serve as somewhat of a screenshot that rids the need to explore deeper. They give the surfer a shallow preview, making arrival at pages more rhetoric or unneeded. The surfer can become selective based on the appearance of the destination page.

Screenshot-type extensions turn out to exist in Opera and even in Mozilla Firefox. In fact, there are several such extensions for Firefox. One of these extensions is BetterSearch, which manipulates search engine results pages directly. Thumbnail zooming is possible too and thumbnails rely on Alexa’s thumbnails database with the caveat that only front pages are displayed, rather than the actual pages delivered to the SERP. In BetterSearch, almost a dozen search engines are supported. One has to remember and constantly bear in mind that a fast connection is required for pragmatic use. It can otherwise become an obstruction — a time-consuming obstacle

I choose Firefox primarily because of its
extensions. Some site continue to snub it.

ROWSER discrimination, as it is frequently referred to, is always a bad idea. Sadly, many sites still use deprecated JavaScript code to distinguish between Netscape and Internet Explorer. In reality, most of today’s Web sites can be rendered almost ‘pixel perfect’ by Camino, Safari, Mozilla, Opera, Netscape, Firefox, Internet Explorer and several more obscure Web browsers, even Konqueror.

Vis-a-vis Konqueror, functionality in Konqueror is being extended all the time and its renderer (KHTML) improves rather quickly too. It no longer plays catch-up, but the acceptance of new browser is hindered by vain user-agent sniffing, which in itself is a terrible practice. In the case where proxies are being used, the assumptions might be wrong. In other circumstances, this promotes agent forging/spoofing, which only adds noise to Web statistics, adds bloat the applications and makes an unhealthy industrial environment for all to compete in.

This is a testimony which I could not help quoting. It speaks of the rotten state of Web design courses.

I asked the teacher to direct me to a real website in the real world that he has developed. He told me his job is to teach, not create a site. I then asked if he could accurately describe the use of the [a href] tag and how to make it open in a new window. Total silence. I then asked if he knew what “CSS” stood for and its place in web design. Once again, silence. I asked what PHP is. I asked about asp and its purpose in web design. Still silence. I was sure he hung up on me, but I was wrong.

After 10 seconds of stammering his way through the css question, he summarized his comments by telling me that CSS is a very specific programming language seldom seen in the real world. He admitted he was not familiar with asp and phc (that’s right, ph”c”). He feels that since it is seldom used, there is no reason to teach it to the students.

IREFOX keeps eating Internet Explorer. Not literally so, but this Web site is solid proof of the growth of Mozilla Firefox. Below are browser statistics which have been accumulated for over a year and embody nearly 100 gigabytes of traffic.

Generated for Analog