Month: May 2004

Bottom line is I don’t normally like frameworks. But I like Tapestry.

This is what I typically hear from people who use Tapestry. It’s an awesome flexible Java application framework which uses the concept of reusable componenets throughout the entire system. It scales well, deals with internationalisation easily, has masses of code reuse because of the components and has great error reporting. Some people have compared the way it works with the fundamentals behind Swing, and they’re not necessarily far wrong. What is true is that its learning curve is much steeper than that of other Java frameworks I’ve used (ok, was, when I was learning it about 18 months ago), but the mailing list support is excellent, there’s an active wiki, and the lead developer, Howard Lewis-Ship has published a book (as well as having a blog), so there’s no real excuse to not at least take a look.

How to make a good ID in Atom by Mark Pilgrim clears up one of the murkiest aspects of Atom for early adopters – how to generate a GUID for items in an Atom feed (well, it was murky for me at least).

As with the the rest of his how-to entries it’s clear, easy to understand and has a worked through example, allowing even dolts like me to get what he’s talking about. If you want to generate or parse Atom, you should read it.

SauceReader recently went to 1.4.1 and with it comes MS Messenger integration (described more fully here but you can get per-user feeds for other people that are using SauceReader and if not then it’ll just send a message with the item title and link), better memory usage and the clearing up of several other issues (check the status on the SauceReader weblog!).

More importantly (hah!), their website now works in Firefox (although still not without its glitches!) and in particular the screenshots page is actually viewable! woohoo!

Interestingly, Scoble just picked up on it and since then its picked up a whole host of links

SauceReader is good, it’s really starting to push the bounds of what we expect from a desktop aggregator (what with the Messenger integration and built-in blogging tool), and is rapidly becoming the answer to what I asked last month., although when I first used it I thought it was just going to become a simple aggregator like SharpReader. I’m glad it hasn’t.

Transvestites in an atom feed (from In Case You Thought The Syndication War Was Entirely Humorless Handwringing)

JTextile 1.1 is out!

There have been a lot of changes in this version, and a massive amount of bug fixing (for the record, it doesn’t suffer from any of Textile4j’s current bugs). It’s hugely reliable, fast, and easy to embed.

It makes use of the GNU regexp library so it can be used on 1.3 JVMs as well as 1.4 and above.

All-in-all, it’s awesome. Get it now! 🙂

Is it just me, or does RedHat’s Bugzilla Install put Mozilla’s to shame?

Yes it does. Bugzilla’s user interface, particularly the query page and individual bug pages are incredibly difficult to use until you’re used to them. There’s a real usability problem there.

If I ever have to use Bugzilla, I use Bugxula instead because it’s just so much easier.

I’ve just come across the CSS Property Index which is an alphabetical listing of CSS properties. Of course, there are many of these of different kinds across the web, but I’ve never found any as useful as the DevEdge Netscape Sidebar Tabs which provide Mozilla sidebars for CSS 2, CSS 2.1, HTML 4.01, XSLT 1.0 as well as DOM 2 reference and others.

The sidebars will also work under Firefox, but with several caveats. Firstly, because Firefox doesn’t implement sidebars in the same way as Mozilla, you can only view them in the sidebar if you bookmark the URL of the sidebar itself, and then check the “Load this bookmark in the sidebar” box in the bookmark’s properties window.

Secondly, because Firefox doesn’t have a sidebar-proper, you can’t just press F9 to see it, as in Mozilla – the easiest way I’ve found is to move the bookmark into your Bookmarks Toolbar Folder, so that it appears in the toolbar, within easy reach.

Thirdly and finally, the references provide a tabbed interface, which in the CSS reference, for example, breaks up the different media types properties (visual, aural, paged), and provides access to the spec TOC, and how to use selectors, except that when you click them in Firefox, they open in the main browser window instead of the sidebar, which is where they’re supposed to be.

Nevertheless, despite all these annoyances, they’ve remained in my toolbar for the past 18 months, and aren’t likely to disappear any time soon, because when it comes down to it, all you want is a quick and easy, usable, navigable link to the spec, which is exactly what these provide.

The URLs you need to bookmark are: CSS 2 Reference and HTML 4.01 Reference.

One of the many sites I read linked to findability.org today. The thing is, whilst the people behind it may know a lot about findability, I’ll never know because I can’t read their site. A nice light-grey text on a white background gives me so much trouble that I just won’t read it. I could spend some time customising their site with my own CSS, but frankly, it’s more effort than it’s worth. My time is valuable, and if you’re not going to design yore site in such a way that I can read and navigate it effortlessly, then not only am I not going to stay on your site, I’m not even going to skim read it (other than to look for a style switcher).

Didier Hilhorst recently wrote When Black Is Not Black, a piece on why you shouldn’t use black for website copy. It’s obviously a little tongue-in-cheek, and reveals an awful lot about designers (especially one so noted for his CSS skills), but is very telling. The thing I care about more than any other in design is that I can read the text. Really. I’m an end user, I’m no design genius. I won’t be told that the contrast on my monitor is all wrong (which has happened before), because frankly, everything on my computer except your site looks just great thank you very much.

The thing is, of course, that most designers hate black-on-white text. OK, I’m generalising, but look at the list of the 80/20 design blogosphere and see how many use black on white text. Finished? Not many, was it? Of course the thing with legibility isn’t just about the choice of text and background colour; the font, its size and the line-spacing all contribute (typography in general deals with many other factors but most of these just aren’t available for change on the web). For example both stopdesign and Simplebits have extremely legible sites with white background and non-black text (without looking it at their CSS, I’m guessing at about #666), mainly due to the large text and line-spacing, much larger, in fact, than you’ll typically see on a commercial website. Think of a few large global organisations, check out their websites, and see what they do. I looked at Microsoft, Dell and TimeWarner – they all have pretty similar text sizes, but the TimeWarner site is far less legible to my eyes. Checking the styles, it uses #333 for text.

Now of course, this isn’t meant to be conclusive proof of, well, anything really, but as long as your text remains below a certain size (let’s say 0.75em with 1em line spacing for an arbitrary line in the sand) you’ll most likely be better off with black text rather than something “close enough” like #333 or #666.

Actually, Didier has followed up his previous post with Contrasting Vision which offers his readers the chance to select the most readable copy text and constant background colour and let him know. Of course, it took me quite a while to get there, because his #4C4C4C text makes my eyes blur. 😉