Monthly Archives: February 2006

RAID Array Management Interfaces

Posted on by

There is a huge problem with every RAID array management interface I have ever used, they don’t provide many hints as to what’s actually going on. For example, today I was trying to mirror an existing drive on an IBM server we have here. When I went into the management interface, it would have let me create an array (with the existing hard drive) however nowhere was there any help text or hints about whether or not that would blow away the data on the drive. I believe the work it used was “create” and there was in the same menu an “init” option. Whoever designs these interfaces needs to learn that intuitive design isn’t only good for non-technical users.

Good UI helps everyone.

Clubs and Open-source

Posted on by

Franck Hecker, Gervase Markham and Chris Blizzard have all been talking about ways to get more people involved in open-source. What I find interesting about what they are saying is how similar it is to any other sort of volunteer work. Sporting clubs and charities around the world have been asking these same questions for ever. How can we get our members, to be volunteers? or How can we get our users to become contributors?

To me, one of the best ways to get and keep people involved is to lower the barriers of entry. If someone comes along to Bugzilla and tries to find out exactly how people would like the bug fixed and can’t get an answer, then they may just move on to the next thing. The same as if you want someone to work the canteen at your football match and they have to fill out 3 insurance forms, they will most of the time, forget about it.

Being a good mentor doesn’t necessarily mean dumbing down the work, but providing a lot more support along the way. Mozilla is starting to do this with devMo, but sometimes someone doing a bit of personal hand-holding is the best way. This carries over to Frank’s sales metaphor perfectly, by working harder to close the deal.

Mentoring involves telling people what to do, the same way you would teach a new employee how to do things. This means, rather than telling them to just jump in there should be a list of bugs that are minor, lower priority, low risk changes that someone can learn the codebase from, and receive feedback. Feedback in the form of advice on how to tackle the problem, landing the patch, and review.

As far as I can tell, the current state of Review is false economy. I’m aware that there aren’t enough reviewers to manage the all of the incoming patches. However without them being reviewed, new reviewers or engineers aren’t getting the aformentioned feedback.

Google Reader, Mail and Talk

Posted on by

Recently I changed my email and RSS subscriptions from Thunderbird to GMail and Google Reader. I’ve been really happy with how most of this works, except for the default from address which isn’t a show-stopper but it is annoying. So I’m pretty happy with the inclusion of Gtalk in GMail. I tested it a bit this afternoon and now all I need is some of my friends to use XMPP.

The other cool feature of Google Reader is it’s built-in podcast player. Basically if there is an enclosure on a post, it will present you with a player bar and you can listen to the mp3 right there in the webpage. This was a really welcome surprise when James recorded a couple of guitar tunes.

Russell Beattie’s Absence

Posted on by

I’ve been contemplating writing a review for my mobile phone, (which I’m not particularly happy with) and I realized that Russel Beattie hadn’t been showing up in my aggregator (google reader) for quite some time. I knew he was going on hiatus across Christmas but thought he would be back by now.

Turns out he’s been writing since January 1st and he has turned off his comments which turned out to be a bit of a storm in a tea cup. Anyhow, it turns out I was subscribed to rss.jsp on his site, and it doesn’t appear to have a redirect on it.

Regarding the comments Russ, good work taking the plunge and trying to get your shit sorted. But remember it’s nice for your readers to be able to see who is linking to you. Maybe some links to per-post Technorati pages would do the trick.

Waferbaby Interview

Posted on by

I’ve been meaning to start to interview some personalities of the Internet, especially Australians and post them here. For various reasons, but I think that often you get the perspective of people that they want to show, rather than what you want to know. (Poetry!)

The first of these interviews is Daniel Bogan aka. Waferbaby. Waferbaby is a weblog-cum-community that has been around pretty much forever. Daniel works for an Australian hosting company, Segpub doesn’t eat meat and draws some pretty cool sketches. Anyhow, to learn more check the website or read on.

Tell us about how you got to where you are, with work, especially Segpub. What helped you get work as a codemonkey in Australia?

Trial and error, really; I never had any serious web education, just jumped in the deep end and started to work out how things were glued together.

I do remember that in early 1995, the multimedia (remember that?) company I worked for sent me and their graphic designer off to a one-day HTML course at Apple Australia; this as around the time that the first WYSIWYG markup editors were popping onto the scene. I remember that the teacher’s name was Spider, and not a lot else.

What about the more programming side of things, being self taught, did you just figure it out slowly or did you get thrown in the deep end and be trialled by fire?

Usually the latter; even today that’s still how I tend to do things.

Do you have delusions (aspirations) of grandeur with waferbaby, or are you happy with it being your little baby, mainly a blog with some cool stuff attached?

I’m very happy as it is, which is surprising; it’s taken me this long to really become settled with the site (which explains why I’ve torn it down and rebuilt it from scratch so many times).

I see it as a constantly evolving application, basically – something I can keep playing with and adding to while I potter away with Real Work(TM).

Tell us about any crazy new plans you have in the pipeline for waferbaby.

I’ve always got these huge, crazy ideas for the site, but whether I’ll find the time/inclination to implement them remains to be seen (read: probably not any time soon).

While you don’t call yourself a designer, waferbaby always looks awesome. What are some of your major influences when building websites?

I always prefer to let the content speak, so when I go in to tear the stylesheet apart and start again, that’s at the forefront of my mind.

But seriously, I’m so not a designer; I’ve really only had the one design in various incarnations and colour changes over the years!

Well, I think you do a pretty good job.

And I think you’re deluded. 😉

You’ve got a pretty awesome relationship with internet bigwigs like Jeffrey Zeldman, John Gruber and my one of my personal favourite writers Dan Benjamin. How did you come to be friends with these guys (living in Australia)? Has this helped your work?

Just one of those things that happened organically, and by chance.

Hell, Jeff emailed me out of the blue about six or so years ago telling me he liked the site, and at the time, I had no idea who the hell he was! So I thanked him, we just started chatting via email, and I’ve been bugging him ever since.

I can’t actually remember how I came to be friends with Grubes and Dan, actually. Oops!

Regardless, it’s certainly helped me career-wise, since I’ve been given the gift of working with some of these talented bastards.

You’ve joined the growing list of people that have moved over to using Rails, hows that going for you? I’ve had a look at using it, and it looks pretty sweet, but tell me about some things about it that you don’t like, or that you would do differently with the framework if you could.

Nothing – if it ain’t broke (and it ain’t, in my experience with it) then don’t fix it.

What about PHP, what were the reasons you moved away from it? Mainly for a change? or were there other things that pushed you?

I’d actually moved away from PHP a while ago – the way that the OO parts were kinda thrown together in a mess got on my nerves.

I’d switched to Python and started building my own Web Framework (because this is what people who use Python do, apparently) with a bunch of friends, but all the while I kept hearing people praising Rails, so I gave it more than a cursory glance this time around.

Eventually, one thing led to another and here we are.

Where did the tagline, “We eat bandwidth for breakfast” originate from?

Oh god, who knows? It’s about as old as the site is, so I can’t even remember; something I pulled out of my ass randomly, I imagine.

I mean, that’s how everything else came into being.

And in the spirit of one of these meme’s that spread like wildfire, have you got a couple of websites that the readers probably won’t have visited that you would like to share?

You know, I actually don’t.

Sad, but my reading (in NetNewsWire) is quite vanilla and boring – Mac news, video game news, general tech news, a few webcomics and those of my friends who bother to keep their sites updated regularly.

Which isn’t as many as it used to be, sadly. You bastards!

Date Interfaces

Posted on by

Creating the interface (UI and implementation) for a date is a tricky process. The tradeoff between ease of use and speed is hard. The airline websites often use a dropdown calendar, however that doesn’t work if the form is being used for data entry. Alternatively, a text box that may have some help text doesn’t assist the user much at all, leaving questions such as do I use dashes or slashes. Month or date first? This option isn’t very learnable as a user can’t tell quickly that a date needs to be entered here.

Ideally a combination of the both would be the solution, as sometimes is seen where there is a text box that can then be dropped down to calendar. This however isn’t exactly easy to replicate for the web while still taking into account things like tab order.

Another consideration is when the date has a range, ie. In the past. This user interface should show that future date’s are unacceptable, but at the same time not frustrating users.

As a side note, an interesting way to tell how difficult a solution for an interface is, is to look at how many different implementations exist.