Category Archives: tech

Weekly Head Voices #9: Windows 7 Geek-o-Rama.

I’ve unfortunately not been involved in any quantum entanglement accidents recently — teaching duties are mostly to be blamed for my two-week silence.  Besides spending at least a whole work-day every week on our Data Visualisation practical, I’ve been lecturing and also been preparing a new lecture block on information visualisation with a dash of visual data analysis.  Due to my not secretly being an infovis expert, this latter activity has taken up quite a chunk of my time and effort.  On the other hand, the exercise has forced me to acquire a significant amount of new infovis brain juice which I’ll probably soon be applying to impressive effect.

In any case.

On the geek front, I have to say that I’m really liking Windows 7.  Partly eye-candy, partly the SuperBar, partly the revamped file manager: I’m a happy camper both on the NetBook (just installed 4GB of SDHC especially for ReadyBoost) and on the quad core workstation.  Feel free to discuss this in the comments, even, or especially, my black turtleneck-wearing friends! :)  Unrelated to the big 7, it turns out that if you use large removable (USB) drives between computers with different operating systems, NTFS is your best bet.  Even more unrelated, I’ve also discovered that one can implement complete independent Windows applications using AutoHotkey.  Before I knew what I was doing, I had re-implemented most of my envedit application in AutoHotKey, with GUI and all (you can find my efforts in SVN, AHK’s a strange little language).  The resultant stand-alone app is 400k, which compares favourably to the 4MB envedit installation.  To conclude this week’s edition of I Really Like Geeking Out, I broke down and bought 20G of extra Google storage for slightly less than EUR 5 per year.  I’m not using it (yet), but I somehow get a kick from seeing this at the bottom of my GMail interface (click the image for a slightly larger version):

gmail_storage_screenie

With regard to research, things have been going just swimmingly.  There are a number of really cool articles being lovingly incubated as we speak.  Some mathematical visualization and some time-varying VDA will go to Eurovis, whilst one other submission is already being carefully groomed for the Vis deadline in March next year.

In other news that absolutely made my day on Monday, November 9, 2009: After being in the oven for almost two years, our pathological shoulder segmentation article should soon appear at a news-stand near you:

Peter R Krekel, Edward R Valstar, Frits H Post, Piet M Rozing, and Charl P Botha. 2009. Combined Surface and Volume Processing for Fused Joint Segmentation. The International Journal for Computer Assisted Radiology and Surgery.

Mr Cricket, that was just marvelous!

Finally, I’m ecstatic to report that due to an unfair dose of serendipity, not in the least brought about by the involvement of one extremely resourceful individual, Longitudinal Medical Visualisation (google should take you to the right place) looks like it might be getting off the ground in a Really Big Way soon.  Stay tuned kids, stay tuned.

Weekly Head Voices #8: Uninterruptible Fun Supply

Dear readers,

Due to a small accident with a friend’s quantum entanglement device, I briefly got stuck in a high pressure reality vortex. The headaches have subsided, but I do still seem to be suffering from slight time compression artifacts. In any case, that’s why there’s only this one edition of the Weekly Head Voices to cover weeks 43 to 45. As is always the case, please make use of the bolded phrases to guide you through this post. In other words, the fat words tell you what you you might find interesting so that you can skip the rest.

Week 43 was for a large part about re-learning a lesson that I’ve learned and forgotten more times than I care to count, but it was mostly about joining Superbly Cool Extraordinarily Lovely People (hi there y’all!) and going here:

YouTube Preview Image HEAD ASPLODE!

Now how about that lesson? Let’s go:

On the Importance of Not Getting Interrupted.

During these past weeks I’ve been hard at work completing a mini-thesis (some call it a teaching portfolio) documenting my teaching activities, meaning that I had to spend a significant amount of Contiguous Time(tm) producing a significant body of text. In order to supplement the scarce supply of said Contiguous Time, I spent two mornings working at home. Furthermore, I for some or other vague reason decided not to check email before I started early in the morning and of course also not to keep my e-mail client running whilst working.

My word, what a difference!

Who woulda thunk it, it turns out that that habitual and reflexive email checking really breaks one’s speed and, in my case, causes unnecessary stress as each time the inbox piles up with even more remotely injected work. Bottom line: I’m going back (for the umpteenth time) to 3 fixed email checks and inbox emptying sessions per day: one before the early morning daily review, one just after lunch when my brain is too busy coping with digestion anyway and one in the late afternoon.

Operating Systems all-you-can-eat Buffet

During wind-down time in these past three weeks, I installed and tried out the following operating systems:

  • Moblin 2.1 preview on my netbook: Oh my it boots really fast and is very pretty. It would take some getting used to, my experience was too much mobile internet device and too little computer.
  • OSX 10.5.7 on my Q9450 quad-core: First: No, I have no idea how that got there! Second: Meh. Looks nice, not my thing though. Third: Eventually I’m going to port DeVIDE to OSX, when either wxCocoa or pySide is ready. I’m only doing this for my goateed, turtle-necked and beret-wearing apologist friends and definitely not for the OS or the company behind it.
  • Ubuntu 9.10 Netbook Remix on my, err, netbook: My jotted down thoughts at the time: Very slick, clutter interface (including maximus) is great for netbooks. It seems the ath9k wlan adapter still has minor problems connecting / staying connected at full speed.
  • Windows 7 on my netbook: Yes, the TU does in fact give us all licenses for this type of stuff, it’s a cool perk. Wow, it went on there quite easily, I simply ran “setup.exe” from the unpacked ISO and installed it to an extra 70G partition. After installing the usual suspects (truecrypt, 7zip, avira, vim, asus stuff [Super Hybrid Engine, Hotkey, Asus Update, Touchpad driver], fastcopy, chrome), I was up and running. Looooong battery life seems to be intact.

The End, My Friend

In week 44 a number of us went to defend the whole TU Delft Computer Science research programme at an international research evaluation. Besides leading to my recent PowerPoint post, this occasion surprisingly turned out to be great fun (probably thanks to the 5-star evaluation committee and their interviewing style) and we seem to have done quite well in the evaluation.

Preparing for the evaluation and finishing my teaching portfolio took up much of my time, so much so that I have not been giving the people around me all the time and attention that they deserve. People around me, I am acutely aware of this and I will make it up to you!

Your GMail account CAN be hacked over insecure WiFi

Today The Next Web posted an episode of BBC Watchdog where it was demonstrated how a GMail account was hacked through insecure (WEP) WiFi.

https_gmail_url

For those of you still wondering, I’d like to confirm that it is indeed possible to hack a GMail account over insecure WiFi: GMail does indeed always send your password through secure HTTP (SSL) so that this can’t be directly hacked, BUT, by default, the rest of your session happens through normal clear-text HTTP.  The Watchdog episode of course gives absolutely no technical details, but it’s most probably the “sidejacking” attack first published by Robert Graham, where the attacker reads the cookies of the post-authentication HTTP traffic and uses them to fool GMail into thinking that they are in fact the legitimate owners attacked GMail account.  This attack works on other webmail and -service providers too.

In short, if you EVER use a network connection that you don’t trust, simply change the “http:” in your URL bar to “https:”, or, even better, change your browser connection to “Always use https” on the GMail Settings – General page.   With both of these solutions, the whole connection will use secure HTTPS (SSL), and cookies can’t be sidejacked.

The drawback of the secure setting is that your GMail access will be slightly slower than usual:  The encryption costs more compute time at both ends, and the transmission of data is slightly less efficient.

Two PowerPoint 2007 tricks that could save your life

On Wednesday evening I was putting the finishing touches on probably the most important presentation I’ve given in the past few years.  As I was testing everything on my trusty little netbook just before bed-time, two scary problems reared their ugly heads: 1) An embedded MS-MPEG4 encoded AVI simply showed a black box when played and, perhaps even more disheartening, 2) the last and most important slide took between 4 and 6 seconds to appear.

the people that brought you powerpoint

The people that brought you PowerPoint.

Some of you might remember that I posted about a strange PowerPoint 2003 bug and its work-around back in 2005.  In short, hyperlinking to an MS-MPEG4 encoded AVI would result in a completely unnecessary warning dialogue.  Renaming the AVI to MPG made this problem go away.

Guess how I worked around the movie clip black box problem in PowerPoint 2007?  Yes, children, changing the file extension from AVI to MPG and re-embedding the movie-clip resulted in a perfectly-playing embedded AVI.  You can thank the clever guys and girls at Microsoft for keeping life interesting with these pleasant little surprises.

Problem #2 took slightly longer to solve, but Uncle Google soon had me on the right track.  The last slide in my presentation, the one that took 4 to 6 seconds to appear, was quite image-heavy.  It turns out that PowerPoint saves embedded images in a relatively inefficient non-compressed form.  This does, contrary to what we expect from MS, make sense, as one would like to retain all image information.  However, when one is doing a slide-show, the processing that has to take place scaling and displaying all images takes an inordinate amount of time.  Having an upwards of 4 second delay before the next slide appears is simply not acceptable.

The solution is saving a copy of your presentation, but with image compression activated.  Do “Save As” and on the bottom-left of the dialogue that appears, select the “Tools” menu button, then “Compress Pictures”, then “Options” and finally “Screen (150ppi)”.  Now “Ok” twice, then finally select a new filename and click on “Save”.  In my case, the presentation file was reduced from 10MB to 2MB, and the load time for that crucial last slide from 4 seconds (minimum) to slightly less than 1.

That concludes the second installment in my exciting life as a PowerPoint flunky!

Futuristic Betting at VisWeek 2009.

So I went to IEEE VisWeek 2009, and it was far more awesome and enjoyable than even my most optimistic expectations. Besides contributing to the tweetstorm (see #visweek) but not being able to liveblog due to higher priority activities, attending paper presentations and chatting with as many cool people as possible (much higher priority activities), this year I’ve also made a number of elaborate bets with a subset of said cool people concerning the future of our technology. If all goes according to plan, the bets’ll end up being visionary, if not, they might be slightly embarrassing and we’ll have a good laugh at VisWeek 2019.

Because these bets only realise in 10 or 20 years, I’ve told my betting partners that I would write it up on my blog so that we could check in that much time, and that they would then owe me copious amounts of beer. This also gives them the opportunity to check my wording for suitability, as we might have to cleave hairs when the time comes.

Bet #1: At VisWeek 2019, I and at least one other person will be wearing a HUD pretty much all the time, OR I will have at least one bionic eye.

Courtesy of Fraunhofer-Gesellschaft.

Courtesy of Fraunhofer-Gesellschaft.

I made this bet with the infamous Dr. Bob Laramee, proprietor of the 5-star webpage.  This came up in a conversation about having some form of non-intrusive display device always available with which things could be looked up, relatively unobtrusively, during for example a conversation.

I do expect that sometime within the next 10 years, HUDs (heads-up display) will be offered as an option with every new set of spectacles.  This is definitely not a push-over bet, but that makes it all the more exciting.

I’m not sure how the bionic eye clause slipped in there, honestly!

Bet #2: 19.5 years from now, there will be more cars with alternative propulsion systems than there are cars with fossil-fuel based internal combustion engines.

This one was made with Dr Helmut Doleisch, linked view data analysis guru and now CEO of SimVis (hey man, BUY THEIR SOFTWARE!).  He is concerned with all the vested interests in fossil-fuel combustion, whereas I think that 19.5 years is more than enough time to introduce non-fossil fuel based propulsion systems on the road, so much so that less than 50% of cars on the road will have need of fossil fuel.

Brazil with its 190 million inhabitants is an interesting example in this regard: A large number of their cars already run on either gasoline or sugarcane ethanol – clearly a great step in the direction of non-fossil fuel options.

Bet #3: In 2029, distributed conferences with tele-presence will be common.

In this case, the beer-donor is Dr Stefan Bruckner, master of all things volume visualisation and father of the VolumeShop software system.  My contention is that in 20 years, conferences where attendees take part via tele-presence systems, whatever they may be, will be common-place.

I’ve called one of the models I envision of this “clustered telepresence”, which would entail that groups of attendees would indeed gather physically, but that these remote clusters would be connected by advanced tele-presence systems, involving advanced displays (think very large, or perhaps even some form of mobile volumetric displays) and distributed and mobile sound.

During the discussion preceding the making of this bet, valid concerns were raised with regard to the efficacy of remote socialising, especially beer-drinking.  I agree that this is an issue of utmost importance, but still contend that technology and efficiency concerns will conspire to address this problem in a way that is at least good enough to fool, to a sufficient extent, all parties involved.