Simon Rumble’s blog

The marketing gurus at my work want to redesign our online ordering system. What we ended up with is certainly problematic, but at the time was about as much as we could get into the scope of this huge project. It makes sense, now that the dust has settled, to revisit and massively improve it.

The trouble is, there were a lot of thought that went into the specific flow we finally got. A lot of complex, important decisions were made, and they haven't necessarily been documented.

To help in our design process, I suggested we go straight to proper mockups, skipping the wireframe stage since we're pretty clear on what we want. Now I'm digging around for the best way to do this.

The mockups need to have the majority of the form logic we'll be using, as the form logic is the most important part of our improvements. We need to shake out the dependent questions we'll need to ask, so we can structure it all properly.

So, dear lazyweb, does anyone have any tips for tools or techniques to use in mockups? These need to be more than just "white site" type stuff, though styling isn't anything like as important as the underlying logic.

Jeremy, shamelessly pimping for his work has suggested using a PS3 as my MythTV front-end as an alternative to building my own front-end. Advantages being it could be slightly cheaper (depends on hardware choices), Blu-Ray drive thrown in, and it's got some pretty awesome hardware. I'd add that I'd get a PS3 as a significant advantage.

It's a very tempting idea, and one I'll think about quite seriously. Current MythTV support seems a bit patchy, but it also seems to be improving pretty quickly. Games consoles are a natural platform for this kind of hardware: it's already designed for the lounge room, it's got all the hardware you need already, and geeks wouldn't complain about a shiny new games console in the lounge room.

I might see how I go with SD and the new telly for a while and keep an eye on progress of MythTV with the PS3. If I decide I want HD, it's a very good option.

Since the FFA seems too clueless to provide this, I've created a Google Calendar with all upcoming Socceroos matches. I watch the football news about our team pretty closely, so I should be able to keep it updated. Times are in Sydney time, though your calendar app should be able to translate that for you.

Google calendars can also be imported into other apps. XML, ICAL HTML.

I've had quite a bit of response to yesterday's post about HD MythTV. Thanks for the input folks, you've pretty much answered all my doubts!

Cafuego told me he's having no trouble recording two HD streams over an NFS share on 100 meg ethernet, so local SATA shouldn't have any trouble. That's about what I'd expect too, though commercial flagging crunches the drive a lot. Spread between two drives with the new Storage Groups feature, I reckon it'll be fine.

Dave wrote to tell me he's happily watching HD on an AMD64 3000+ CPU and an Nvidia GPU. He points out that XvMC acceleration of MPEG decoding is important, which is the same as I've found with my silent SD front-end. One thing I've been wondering about is whether Australia's DVB-t solely uses MPEG or whether one of the other standards is also in use. The Wikipedia page mentions use of the H.264/MPEG-4 AVC codecs, though I suspect that's a very new thing and not used here.

Dave also points out that [wife] "Rozie never seems to get around to watching half the stuff she schedules", but that's pretty much the point of a PVR for me. Record everything that might, possibly be interested, but don't get too upset if it expires off unwatched. The occasional night you spend in wanting to blob out with lots of telly makes it worthwhile. The approach becomes "what's the telly got for me" rather than the old style "what's on telly".

Matt chimed in with some great info about the Panasonic LCDs I'm looking at. "I helped my parents pick out the 700A a couple months ago, and they (and I, when I'm around) have been really, really happy with it. The picture is extraordinary, and the signal processing (noise reduction, motion comp.) it does on the (crappy) foxtel and (sometimes crappy) free-to-air signals is amazing. Better than far, far more expensive 42 and 50 inch plasmas I've seen." Glad to hear it and that's been my impression from various forums and reviews around the traps. Not so sure about the motion compensation and noise reduction which seems to be the main difference between the 70A and 700A. Will have to check it out in a shop, side-by-side, I think.

Matt also suggest looking at Via Mini-ITX options which "are fanless and feature a DVI port". My SD front-end is an ME6000 and it's brilliant. Absolutely silent, which is ideal for the lounge room. The problem is, the newer EPIA boards that might have the grunt for HD look great, but I haven't been able to definitively confirm that the free drivers support the new chipsets, or that there's enough grunt for a 1080i stream. Anyone able to shed any light there?

TimC had some suggestions about fixing the green tinges on the old CRT. The green tinge appeared after we moved, and could well be due to proximity to speakers. He's suggested trying some degaussing or magnet actions. I might give it a go just out of curiosity, but the other half has agreed to an LCD so I'll still go down that path ;)

Craig described his VDR/Xbox running XBMC setup which sounds quite cool, but it wouldn't be up to HD and XBMC doesn't integrate with MythTV particularly. His comment about HD broadcasting in general rings pretty true though: "personally, i think that unless your TV is fantastic then you just aren't going to notice any difference between HD and SD except that the former takes up a *LOT* more disk space." I'm inclined to agree there, particularly if you haven't got a 1080i/p screen. 768 lines of resolution is certainly more than SD's 480, but it's not really that much.

I think I'll take his advice and see whether I care enough to upgrade from SD rather than dive straight into it. I suspect the improvement in quality won't be all that important, and the improvement just from having a bigger, better telly will be enough. For now, anyway.

So thanks for the responses folks. Most enlightening!

Our ancient CRT telly has recently started getting a green tint in the lower corner, and Holly just got a substantial pay rise by moving jobs, so we're going to treat ourselves to a new telly. Judging from what I've seen in forums and reviews, the Panasonic 32 inch TX-32LXD70A LCD looks like the business, or possibly the 700A variant if I can work out what the substantive difference is (besides the marketroid-speak "Motion Picture Pro : 100Hz Technology" whatever the hell that is). So it looks like we'll be going the high-def route, not the ridiculous 1080 lines level of HD, but at least somewhat better.

There's a few issues this brings up. First, my front-end is a completely silent box that won't have the grunt to decode low-end HD, let alone the 1080i the commercial channels are pumping out. So that'll need to be replaced, and I'll probably end up with something that has a fan unfortunately.

On the back-end, there shouldn't be much impact. DVB-t capture cards can capture anything pumped out by the broadcasters, including the HD channels. Data throughput could become an issue, and I might do some testing of recording a few simultaneous HD transmissions.

Disk space could become a problem, though I doubt it. I've currently got 600 gigs across two SATA drives. We don't record all that much telly, and mostly on the (lower bandwidth) ABC and SBS. At SD disk space usage, we end up with programmes we haven't watched from months ago being expired off to make space for new programmes. Never been a problem.

It could be that I need to upgrade the network to gigabit ethernet to handle the top-end streams, though I doubt it. I haven't cabled up the new house yet anyway, and I was planning to run Cat 6 anyway, so I should be okay if I need to upgrade.

So it seems as well as the telly itself, I'll have to look into a front-end system quiet enough for the lounge room, but with the playback grunt to decode 1080i (even though I can't view at that resolution).

Anyone got any recommendations or experiences to share, particularly Australian experience?

We are at last on the path to reconcilliation with Prime Minister Rudd saying sorry to the Stolen Generations, nearly 11 years after Bringing Them Home. It's a momentous step and Rudd's speach was magnanimous, stately, appropriate.

Of course the cultural warriors on the other side of the House couldn't resist adding a sour note as their political enemies did what they couldn't bring themselves to do in 11 years of power. Nelson's speech attempted to place reservations on what was an unreserved apology. Abbott couldn't bring himself to clap during the standing ovation. It would have been entertaining to see Howard's sour face as the apology was delivered.

We now have much work ahead of us as a nation to build on this step, but it's a good start.

The product manager for Internet Explorer and his entire team will be the first up against the wall.

My GP has moved practices and is now based in Drummoyne, which is too difficult travel-wise for my tastes. So after being very happy with her, I'll have to change doctors. I'm sure the only industry whose union is also its regulatory body wouldn't allow league tables with qualitative measurements, so I'll have to rely on recommendations.

RateMDs has some reviews, but they're anonymous, don't have many ratings and aren't particularly authoritative. I'd rather solicit reviews where I can at least find out who's giving the review.

So can anyone recommend a GP in the Inner West of Sydney? I want one who gives good advice, is thorough and conveniently located. Not charging the earth would be nice, but I guess if they're particularly good I can afford it.

In response to yesterday's rant about WebTrends, I got an email this morning from Scott Driscoll, Product Manager at WebTrends. He says they're working on improving usability and want a chat about what I've experienced. A very positive sign.

Building software for a general market is hard. You have so many conflicting demands, and the people writing specs and cutting code are generally the furthest away from the people using the software and so have to make decisions without fully understanding what's needed. Then there's the pressure to cram ever more features in. I know, I've been there.

This is a very positive move from WebTrends. Engaging people who use the product every day can only improve things, and I imagine there is quite a bit of market pressure on companies like WebTrends (we're talking figures with five zeros on them to buy WebTrends) competing against Google's free tool.

I'm quite looking forward to giving my feedback. Hopefully I can be quite helpful for them, and the result should at least make part of my day-to-day job easier.

My work uses WebTrends to track people using our sites. It's not cheap, and the user interface is mind bogglingly stupid. It's, of course, one of these JavaScript-heavy, wait-for-everything-to-load-before-clicking-anything-or-it-all-goes-kaboom boondoggles. But that's not the worst of it.

Let's say I want to create a new report. You'd think it would be easy, wouldn't you? Well now that I've done it dozens of times, I nearly always remember all the steps. But let me take you through those oh-so-intuitive steps.

Step 1: Create the report. You think with navigation headings of "Reports & Profiles", "Report Configuration", "Report Designer", "Options", the place to look would be "Report Designer", since what you want to do is, like, design a new report. Wrong. No, you have to go to "Reporting Configuration", expand "Custom Reports", then click on "Reports". "Report Designer" doesn't, it seems, contain anything to do with designing reports. No, it's all about placing reports you've designed into "Templates", which is the next step.

So you design your report. The interface here isn't especially intuitive, but then we're dealing with quite a bit of complexity, so it's perhaps excusable. WebTrends certainly seems capable of some pretty cool shit, but the easy stuff isn't easy.

Step 2: Place your shiny new report in a Template. Next you need to place the report in a Template so it will actually show up. When saving your report, rather than have a nice, obvious next step to make your report show up, you instead have to save the report, remember its name, then find a template to place it on.

So you go in and choose a template. I'm sure there must be people out there using loads of templates, because that certainly seems to be what they're catering for. Us? We use just the one. So you select that, click the "Content" tab, then find a place in the template to put your report. Be sure to let the whole thing load, or it'll completely crap out with the whole application stopping and then you need to close your browser and start again. You click "Add Report" and it comes up with a list of reports you've never touched, rather than the "Custom Reports" you've just created. Once again, wait for it all to load, then change the view to "Custom Report Library" and select your report.

I bet you think you're done now, and your shiny new report should show up. Think again. No, WebTrends has this concept "Profiles", so as well as different templates you can control access to individual reports. Us, we use just the one, but we're burdened with the complexity of being able to have loads more. So now you need to do yet another step.

Step 3: Add access to the report to your profile. So you click "Reports & Profiles", then your profile, hover over it and choose edit from the dropdown that appears. You want to make this report available from this profile, so logically which top-level menu would you choose from "Analysis", "Reports", "Administration", "Advances", "Change History" and "Summary". If you thought "Reports", go to the back of the class -- you haven't grasped the pathological unfriendliness of this application's UI yet. No, to add a report to the profile, you choose "Advanced", then "Reports", go down the list until you find your report and tick "Enabled". Gah!

So now your report should finally be available, or at least available the next time the reports get run.

Someone at WebTrends really really needs to hire a user interface designer and some user testing. A version that makes the easy things easy while keeping the hard things possible is what they need to work on. Or you could just use Google Analytics, which is a damn sight cheaper and a lot more flexible.

What really doesn't help is that, despite being a hosted application, the WebTrends documentation library doesn't have any idea what version you're running. It presents all sorts of irrelevant documentation to you, to help take your confusion to even higher levels.