HaqUse

Moving

2006-12-01T14:09:00.000+01:00

Hi, it was some time since I've blogged here at blogspot. Well, the reason is simple. I have moved completely to blog.reis.se. It's my own space I can do what I like with it and its easier for me to maintain.

Besides when reading my posts through an rss reader its possible to read and drop comments right there..

I might drop a few lines from time to time here but my main space is now blog.reis.se

Annoying home appliances

2006-06-02T14:32:00.000+02:00

I just moved to a new hose, with new kitchen appliances, washer and drier, timers on some of the outlets an so on. This is not super fancy high-tech but still the existence of the all those appliances annoys me, well not the existence. But they appear to me like an obnoxious kid that always needs the full attention from me. Whenever I finally forget that they exist they call out for me. "Beep, hey, look at me I managed to shut off the power to you coffee maker!" "Beep beep, ohh, ohh, I managed to dry all your clothes... beep beep, come look at the clothes now... beeeeeep beeeep COME RIGHT NOW and look at all the dried clothes!"

Read more...

Dotway talks about WPF on MSDN TV

2006-05-08T07:59:00.000+02:00

A couple of months ago I had the oppertunity to attend an MSDN show insperired session with Johan Lindfors and André Henriksson. This was studio filmed and was released two days ago. How could I miss that? It was there in the feed if I just had gotten around to read it.

Anyway, head over to swedish MSDN TV and grab the broadcast it's "the one about WinFX", view it "live" select the chalk talk section and enjoy.

By the way, if you download the film just to watch my pretty face, I'm not in that one. The chalk talks Seems to be missing from the archive.

BTW, it's me in the middle...

Agile Usability No 2 - AID

2006-04-21T14:11:00.000+02:00

I haven’t been writing about the progress so far but here it comes. One of the major concerns was about TLA for the process. Me and a college of mine discussed it for quite some time and the result was AID. For you that are not familiar with TLA; it stands for Three Letter Acronym and it is a must for any technical whitepaper. The AID stands for Agile Interaction Design. And the TLA, at first a joke turned quite good. AID is not intended as a standalone process, it should be an aid for the agile development process. See my point here? ;)

Read more...

Kill of the menu-toolbar paradigm

2006-04-04T12:50:00.000+02:00

I saw this great article, Tipping the Scale (Why the UI, Part 5), concerning the level of complexity when using the, now, old menu-toolbar paradigm. This has been used and worked well ever since it was introduced in the late 70's and went mainstream during the 80's.

Read more...

MSDN Update - WCF at Dotway in Malmö

2006-03-22T14:48:00.000+01:00

There is an event held at our new office in Malmö. It will be a Microsoft session focusing on windows communication foundation and the speaker is Johan Lindfors. You can find more information about the event at Johan Lindfors blog.

There are limited number of openings so hurry up and register

Read more...

SonyEricsson lost it again

2006-03-20T08:15:00.000+01:00

I got me a SonyEricsson w900i mobile phone. The feature list is as full as ever and as usual they didn't do the integration thing the way it should have been done. I just want to point out one of the stupidest thing here. The side buttons.

Read more ...

MSDN TV

2006-02-24T00:06:00.000+01:00

On Friday the 17th, this month, me and my colleague George at Dotway was invited by Johan Lindfors to talk about WPF and what impact it will have on developers and not to forget, the designers.

Read more ...

Crappy User Experience

2006-02-23T10:07:00.000+01:00

Read about this too common Laptop Experience apart from some fun reading, this is an all to well user experience nowadays. Why can't they just make it work? Don't make me think. Don't waste my time on useless things like this (let me waste it on reading about it instead...)

Read more ...

Express yourself with Expression

2006-01-24T14:27:00.000+01:00

So, finally Microsoft released the expression interactive designer tool to the public. I have played with this for some time now and the power that is provided in this small 6MB package is just plain magic. This is the solution for the really cool UX creation.
Now just be careful, don't repeat the blinking and flashing web syndrome :).

Read more ...

Agile Usability No 1

2006-01-03T08:18:00.000+01:00

The holidays are almost over, time to get started on the New Year. So what do I have in the pipe for usability this first part of the year? You guessed: Agile Usability.

There is a big move in agile development. The flexibility it provides and tight integration with the customers give benefits like early and continuous delivery, stability, and reactivity to changes. However in most agile processes there is one big factor missing - Usability. And without usability the risk of failure is as big as ever.

Read more at blog.reis.se...

Digital TV insanity

2005-12-05T07:48:00.000+01:00

Sweden is shifting to digital TV. The analog broadcasting is going to be fully shut down in 2007. Still there are lots of usability problems to solve here. I will focus on the ones that annoy me the most. More at blog.reis.se ...

Usability of speech

2005-11-14T13:36:00.000+01:00

I was one of the lucky speakers at Øredev, a developer conference in the Malmö/Copenhagen region where I live and work. The subject for this particular seminar happened to be .NET and Windows Presentation Foundation (though the name Avalon has stuck) and I talked together with a colleague of mine. However the issue of this blog-entry will not be the seminar in itself but the usability of such.

Read more on the new HaqUse blog site...

Cold facts

2005-10-29T20:48:00.000+02:00

Getting a cold is not always a bad thing. First day, ok, mostly in bed with a terrible fever and pain in every single joint. But the next day I could mange with the help of some ibuprofen.

Anyway, I finally got the time to fix my computer. Started up one of the PE systems, that can be found round the net. This happened to be BartPE, loaded it with Partition Magic and Symantec Ghost. When it was done I backed up my data and booted of the CD. Started Partition Magic and resized my original partition into two partitions. One for my normal work OS and one to use as a test bed. All was done in just a couple of hours including a ghost of my original partition. So finally the system is in shape to try out some cool stuff. Never again a computer without an extra partition to try new and beta OSes (is that a word?) in.

Set it up with WinXP and used the instructions for getting WPF and CPF setup with VS2005, works wonderfully. Now all I have to do is to make this copy a ghost image and I can switch it in and out without using virtual machines. As soon as beta 2 of Vista arrives I know what to do...

So, maybe I should catch a cold more often, I almost never have time for this sort of things. Just one problem I found though, I intentionally planned on ghosting to and USB drive but the PE system wasn't able to find my USB drive (a 250GB Maxtor one touch). And I wasn't able to find any solution for this.

Surfin the airwaves

2005-10-24T15:50:00.000+02:00

Some time ago I tried out a Vodafone 3G card. I had some criticism regarding how Vodafone was handling their customers. Now I have been trying out what 3 has to offer. I had the opportunity to try out the 3G surfing experience using two of ther phones. The Nokia 6630 and the SonyEricsson Z800i. It may be that there are problems that could be solved with the help from customer support but quite frankly, there shouldn't even be issues with customer support.

Using the SonyEricsson Z800i
Installing went quite well, using the 3 software to set the dialup. After that I was able to use it as a normal dial up connection. It should be possible to install without extra software though.

In all the Z800i performed well, maxed out at about 350kbps when pushing it hard with ftp and other downloads. However, while pushing I discovered it was not able to charge at the same rate draining it. Then again, the power need will follow the load. Using it in "idle" state with only mail updates, normal surfing, RSS, etc will keep the charge. Without external power in "idle" state it will last for about 3h.

While travelling from Helsingborg to Lund (on an Öresundståg) it disconnected about 2-3 times, as expected on those train. The good thing is it reconnected automatically without problems.

Using the Nokia 6630
The installation was more cumbersome, forcing me to install PC Suite to get the connection. I have not tried though; now that I know how it’s done it should be possible to set it up with a normal dialup connection.

The 6630 did not perform as well as the Z800i. I had more bad connections (after some time a connection would fail. Seemingly OK but not data is received). Furthermore, I was not able to get the same throughput for the Nokia but after a few resets it maxed at around 250kbps.

The 6630 seemed to keep up with the drain a bit better. It may be due to the lower throughput and that's not acceptable. Both phones should be tested with cable connections as well, ruling out Bluetooth flaws.

The Nokia was also a bit worse in keeping the connection between Helsingborg and Lund and the annoying bad connection that tricked me into thinking the connection was OK makes this a worse alternative.

In total the power of just connecting the computer and if possible getting rid of third party software makes the phone connection alternatives better than using a 3G card. This setup eliminates the need for extra subscriptions for the laptop. However the best solution might be an extra subscription card coupled with the phone subscription.

The iPod with video hoax

2005-10-23T15:48:00.000+02:00

So, (as Microsofties would start the sentence). What’s with the iPod video? Apple is said to be innovative. But there's nothing innovative in the iPod video. They just tweaked an old iPod to show video. Small screen in what, 4:3 format?, hat happens to display a little video.

That's not a video device. It's a desperate response to what all iPod zealots are shouting about. The least they could have done is place a bigger true wide-screen on the damn thing. Place it sideways and replaced navigation with the smaller iPod nano wheel.

True innovation would be a new space saving navigator, adding the biggest possible 16:9 display (preferable an OLED) and use it sideways instead. And ditching the white colour, white plastic are over-used. Even the i should be ditched for marking the new innovation. The nano was in marketing sense innovative, lose the microdrives and use only solid state memory and discontinuing the iPod micro line.

Apple start your copiers, look at Cowon A2, iRiver U10 or Creative Zen vision. (These companies has been said to copy apple far to long now).

At least the "knockoff" companies understand that a personal media player is not a media vehicle. The display is the focus not the wheel.

Off to Hong Kong

2005-10-01T09:47:00.000+02:00

Going to spend the next two weeks in Hong Kong. Will not bring a single gadget with me for communication :) But I will probably end upp bringing back a few.

Dynamic private assembly paths part II

2005-09-30T14:35:00.000+02:00

Well the solution from yesterday sure works ok. And for now it will have to do, but it's not the best or most robust solution. It has to be reworked in the future for two reasons.

First, AppendPrivatePath is to my knowledge going to be obsolete in .NET 2.0. The solution here is to create a new AppDomain, hand it the paths in the setup and then let this new AppDomain control the execution of the application. This will in the end give about the same solution as before.

The second reason is that it would be much better to have isolated AppDomains for plugins, this way the security is handled separately from the main AppDomain and the stability of the main application is improved. I will have to look into this in the future. This should be done for both for stability and compatibility reasons.

Dynamic private assembly paths

2005-09-29T21:22:00.000+02:00

In my work I stumbled on a problem when it comes to the path that is probed for private assemblies. Normally for developed applications there is not much need to have a deeper directory structure for assemblies. However, this solution has certain assemblies in a plug-in structure. And instead of a flat long list of assemblies that are close to impossible to handle, a sub structure is preferred. The following approaches was what I went trough.

The first approach is to make sure that the namespace in the assembly matches the directory structure. This will work but this solution is not flexible enough. It will probably break the first time someone renames a namespace and forgets to update the deployment directory.

Second approach is to edit the app.config file, sure is quite easy to do. The only thing needed to add is the following snippet:

<runtime>
<assemblybinding xmlns="urn:schemas-microsoft-com:asm.v1">
<probing privatepath="fwk">
<probing privatepath="tools">
</assemblybinding>
</runtime>

But this approach does not give more flexibility than the previous. What's worse is that if incorrect information is entered in app.config it may break it for other plug-ins as well.

The third solution solved my problem. In this solution the path is added dynamically to the AppDomain at runtime. By using the AppendPrivatPath method:

System.AppDomain.CurrentDomain.AppendPrivatePath("fwk")
System.AppDomain.CurrentDomain.AppendPrivatePath("tools")

Or if you like:

System.AppDomain.CurrentDomain.AppendPrivatePath("fwk;tools")

Nailed the installation issues!

2005-09-23T20:14:00.000+02:00

Finally got it to work properly. Tried installing VS2005 yesterday, says it can't install MSXML, well duh, it’s already on the damn machine. Whip out a fresh virtual machine now with Vista (thank god for VMWare). Order of installation as Rob Relyea's blog at Longhorn blogs points out. Use the PDC Goods discs and follow this (however with a twist on point 3) order of installation:

1. OS - Windows XP SP2 or Windows Vista Build 5219

2. WinFx CTP installation - Disc 4, installs .NET 2.0, WPF and WCF

3. VS2005 Beta 2 - Copied from disc 2 (Go down to VS2005/vs/wcu and re-install the MSXML6 parser manually first)

4. WinFX SDK - Disc 5 Installs help documents and samples as well as XamlPad

5. VS Extensions for WinFX - Disc 4. VS Project templates

Finally, was going slightly insane there for a while, the odd thing is the MSXML part. A friend of mine had no problem whatsoever.

A shaky presentation foundation

2005-09-22T18:38:00.000+02:00

What a pack of discs PDC brought. The graphic nerd in me sure had to test the goodies out with presentation foundation. If I only could the damn thing to install! Tried a clean system with only Win XP SP2. For one thing, the WinFX setup includes .NET 2.0 Beta 2 as well as presentation, communication and workflow foundation, included are MSXML 6 as well. After around 10 retries in different versions I gave up on XP. A VMWare Virtual Machine running Windows Vista build 5219 did the trick. Finally got the basic presentation foundation to execute. It now shows the sample app, but virtual is soooo slow.

Next thing is to get Visual Studio and the WinFX SDK to run with VS add-ons. Thank god for VMWare snapshots, at least it’s easy to roll back to a known good version for each step. When I eventually get this to work I will sure tell you the exact order.

(By the way, Dell make crappy machines. Why encapsulate a standard graphics card in it's own drivers, making it impossible to test beta drivers for vista? Gaaaahhhh!)

The interaction foundation

2005-09-20T15:30:00.000+02:00

I had the opportunity to attend PDC last week and it was a real busy week. This was my first PDC and all said there was few news compared to the last. Well it might be true but it was clear that the road onward for Microsoft as a developer platform is clear. The presentation foundation and XAML will support complete de-coupling the work of developer and UI designer and a lot of the time this is probably a good thing. Workflow foundation will enable easier handling of enterprise development. This is probably a good thing but there are a few urgent issues to this.

It all drills down to a simple principle: The user interaction will still be dictated by programmers!

If the workflow is dictated by the programmer it will still reflect the inner works of the data in the system and not the user needs and goals. This is a deadly sin in interaction design. And the workflow foundation will not force interaction design into the system. The users will find that they still work against the system or together with the system in best case. It will not reach the status where the workflow fully supports the user.

The same is true for the user interface. It will still be the programmer who dictates what information is available to user, and what the user can do with the information. Sure, the way the information is presented will differ and there will be full artistic freedom in designing the user interface. However, the separation will not let the developer ignore users. The need for interaction architects is as high as ever. I expect to see some great UIs coming out and a whole lot of really crappy ones too.

The tools are greater than never before, but the need for interaction architects is as vital as ever.

My (screen) resolution

2005-06-21T08:34:00.000+02:00

I Hereby state that the screen resolution is not the browser resolution!

For some time now the web designers have been using statistics in monitor sales and usage to foresee what size the website should use. When the screen resolution rises above 1024x768 it cannot be assumed that the user want to use the full screen for browsing. I quick look around my colleges suggest something completely different. Most of them never use the browser in full-screen mode anymore.

A quick, and very unscientific study, suggest that about 70% uses windowed mode instead. I started to examine my own use after that. At times I took screen shots of my browser and saved. After a while I examined those and found that most of the time I used a 900x900 window for my browser (well 800-900). I will continue to observe this phenomenon for a while. It would be great to see some real statistics about it.

But it may not be that the 900x900 size is the optimum, it could as well be that most sites using a resolution of 800x600 as norm. That would mean that when viewing the width is sufficient when I set it to 800-900 and the height, well I don't like to scroll the window so that explains why I don't stop att 600-700.

So do I have a conclusion for this? Well not exactly, I always suggest using a variable width for the main information area. It allows people to use whatever width they have and want, it annoys me when I'm not allowed to use the screen width I want. At least set maximum required width to no more than 800.

The Vodafone virus

2005-06-13T14:36:00.000+02:00

Branding is a good thing on many occasions. Branded cell phones tend to be cheaper and offer a combination of tools and services. But the branding that Vodafone is doing is close to ridicules. The branding is crippling the phones built in software, rendering the user experience inferior. It locks the user and do not offer much in return.

Why do they have to put the link to Vodafone Live at every single extra button. Whenever you make a mistake and press the wrong button you will end up with another few notches on your bill, a connection and a few k of data is billed. For making a mistake! Wonder how much money they make on this.

And the "offered" services? A bunch of links to material that cost way to much, why pay a lot for a crappy ring tone when the original song can be found on iTunes Music store for 99¢? The providers are as greedy as the music companies. On many occasions you pay double for the ring tone than for the original. If they would just offer some real services! And why removing functionality, just because they have not found a way to make money on the function it should not be withheld from the user.

What does this has to do with usability? When the branding interfere and cripples the user experience it wrecks all the effort phone producers put into producing a good user interface. It has taken companies like Ericsson years to go from disastrous UI design to acceptable and now quite good. It takes only months of Vodafone branding to put it back to acceptable and in some occasion disastrous.

On the upside, I tried out a 3G card from Vodafone, apart from the ridicules orange red color and annoying software issues the connection was quite good, but then again credit should go to Fusion for their UMTS Quad GPRS/3G Modem. At first it was a strange feeling to sit down with a latte and be able work, email, read the news and browse. Ok, it wasn't WiFi but it was available at all the places I tried. Now I just wonder how they will be able to brand it harder and force me to use their "services". Hmm… filter my browsing, adding branding to all web pages, force me to use there mail…

NAnt infestation

2005-05-24T08:20:00.000+02:00

I have an urge to automate daunting tasks. Building, compiling documents, etc should be handled automatically. It is far to easy to miss an item, forget to move a file around, etc. This should be automated to lessen the burden on my mind. The command line creeps are not aware of the total time of operation, maybe they are fast as lightning on the basic set of commands they use every day. However, as soon as they are to dig into a new set of commands they are inefficient, slow, makes numerous mistakes, and... Nah, I'm in a good mood, let them have their fun mastering their commands and switches. For now I'm wandering of in the wrong direction here.

So everybody is talking about what a great tool NAnt is and there are possibilities to automatically build binaries, generate documentation from code comments, and put together MSI packages included are also functions to automatically execute NUnit tests. And a lot more of course. Great! Just what I need.

I started to search for information. Downloaded the NAnt software, tried out some builds. Looked promising. However, there is a rather steep learning curve here. At least until you get the idea of NAnt. A new terminology is needed and also the understanding of how NAnt operates. The examples are not good, HelloWorld is not sufficient to get the idea behind NAnt. To be honest the documentation is crappy. Referential only, it fails miserably on communicating the basics thought about NAnt. Attracting new users are certainly not one of their goals. However, bullheaded I continued divining into blogs and articles searching for information.

After finding lots of sites explaining HelloWorldish examples i finally found some useful information. A couple of good links are CoDe magazine and an article at CodeProject. After grasping the basic concepts of targets and tasks I started creating my own build files.

NAnt is quite nice to work with once you get the hang of it. An open source visual "target builder" would be great, handling solutions, projects and properties correct. I decided to set up my NAnt project based on the solution but with my own specification on what projects should be included and what to exclude. There are a few projects from the solution that should be built with special consideration and in a specific order. Finally I got NAnt to generate all the assemblies the way I wanted to. Ok, fine, now some more automation, I want to generate documentation and create MSI packs, I want to automate the testing. Now for the frustration and the failure to fully automate things:

The project I am involved in is using VB, and there are a few things missing in VB. In this case it is the problem with documentation from code comments, etc. There is no simple way to get the XML representing the API documentation from the code (besides using
VBCommenter in Visual Studio). The only thing I came up with was a project named VB.DOC. I won't give you a link as I can't recommend this inferior tool. It has sparse documentation (to put it kind) and in the end it turned out to produce crappy XML with lots of data missing. I hate time wasters like this. Ok, I could live with "manually" exporting to XML from Visual Studio. Then let NAnt build the binaries and documentation using the NDoc task. (There is another option I'm considering; building a command line wrapper for VBCommenter.)

The second point of frustration was MSI packages. Yikes! The documentation for MSI projects is huge and annoyingly complicated. I could as well manually create the packages. To be forced to create the dialogs and interaction manually is far from the visual approach in studio. It's not easy in Visual Studio, especially when you include some extra dialogs and interaction but at least is manageable. The only solutions that would be acceptable to me are either use the .vdproj directly or generate some setup build files with the help of slingshot-like tasks. (Slingshot is a technique where the project files generate NAnt build files that are not depending on sln och proj files).

I read a few blogs getting in to the same problems and some of them gave up, they used devenv to create the MSI from a NAnt exec task. This is not what I desire. Sigh... So for NAnt, I have a few points of considerations:

Do not use NAnt if you want to create API documentation for VB projects!
Do not use NAat if you like to create MSI packages, i.e. if you're not into pain.
Fort the NAnt creators, consider some overview document explaining the goals and basic ideas behind NAnt.

A mess called ClearCase

2005-05-11T08:00:00.000+02:00

I know, I know, it's too long but I couldn't control myself in my frustration. This could easily swell up to a 100-page analysis just touching the surface of the usability issues for this diabolic tool.

A college of mine wrote about a ClearCase experience, a dreadful one that unfortunately is not that uncommon in the ClearCase world. To start with; this tool, or should I say pack of tools, is one of the ~~worst nightmares~~ most complex tools I have come across in a usability sense. Actually it’s not only the usability, but that's another story so I will focus on usability. It was awful the first time I encountered it in 2000 but it was manageable. Now, five years later it has been developed and worked on. I don't know what the Rational developers focused on but it was not usability. They manage to make it an even more ~~disgusting~~ complicated and confusing tool.

When ever I get to lecture on interaction design I use ClearCase. I doesn't matter what problem or flaw I try to elaborate on, I can always find an example in ClearCase. I will not go into details, if I did you would be reading this for days, just point at a few examples of bad behavior. A deeper analysis would probably tell that most problems with ClearCase is usability.

A change and version handling tool should be and aid, it should support the developer with version control and in securing different versions of the code. When it comes to doing this technically I suppose ClearCase does its job, when it comes to doing this as an aid to the user, it fails miserably. The result is 95% overhead and maybe 5% practical aid. It enforces a work-flow on the users that in most cases only add complexity. And now for some usability nightmares:

The dialogs, there must be thousands of dialogs in ClearCase, and they share one common attribute, inconsistency. They are sometime modal and sometimes not. They are fired in multiple, three or four and sometimes more, and then dialogs are fired on top of each other. Often they have lists in them, long lists with no possibility to resize the dialog on the other hand you can get a single text message in a resizable dialog. Clean this mess up, please!

ClearCase don't think much of you as a user. Most of the time it thinks you are an idiot and probably are doing something really stupid. An example is when a file with the same name as a file under source control is dropped on the ClearCase explorer. In my case it was a PDF that should be exchanged for a new one but all I got was the message: "You imbecile, don't you know this can be dangerous? You may lose all your data! You probably are too ignorant to know what you’re doing so I won't let you." I might have cited the message wrong but you see the point. A better way should be to offer to merge the files (well if it's text or code files). Or maybe, just maybe, the user knows what he is doing and let him perform the exchange. After all it is a version handling system, dropping a file in this way should only create a new version of the file the old is kept by the system, right? Be polite!

Error messages are mildly said idiotic, most of the time they give no solution to the problem and you have to search for the answer. There is a term called evil twin. I will not go into the details but there is a solution as well to this problem. This ends up in a text file that you have to close to continue (ohh they do tell you that you should copy the text to a new text editor, they couldn't even trigger a new process for the text...) Reading the text you get an explanation of the steps to solve it, a clear standard procedure that any programmer could solve in code. Why not offer to solve the problem?

I could go on for hours, but it would probably be better if the team behind ClearCase did some work on usability, just the basics you know. Make sure it is consistent, that it supports the user; maybe look at it using some ~~kindergarten~~ standard terminology like REAL (relevance, efficiency, attitude, learnability). At least do something! It can't get any worse, can it? (Well with their current track record it obviously can...)

The only evil twin I could find is ClearCases brother, ClearQuest. But I won’t enter that path, it will probably only get me even more frustrated. Just starting the app makes me want me to defenestrate my computer.

Unusable usability

2005-04-29T15:05:00.000+02:00

This is not an entry regarding development. It's an annoyance in Windows XP. There is a usability feature in XP that's quite good - if only it worked. Whenever you select a registered movie file or film clip the OS tries to show you a preview of the file. Nice feature and it sure is usable. The problem is that for example AVI is not a single type of coded files. It only stands for Audio Video Interleave and is actually a wrapper format for the coded video and audio signal. The coding could be anything the system can handle, so if you add H.264 support it may as well be an AVI file.

Well often the codes are not fully functional or the system cannot interpret the files this means a few things may happen; Explorer may take up most of the system resources trying to figure out the file or it may as well crash and restart. I had problem with the latter, I didn't even have time to rename the file to something other, say mp4v for example (well it worked fine from CMD of course). It was so bad that it crashed whenever I entered a directory with MP4V coded files.

You shouldn't only blame Microsoft for this one, it's sloppy decoders as well but at least there should be a watchdog, timeout or any other fallback functionality. If it takes more than half a second to show a preview then it probably is best to stop trying. And of course it should not crash the explorer.

Well the best thing to do is actually to shut the function of altogether, there is a way: you can use a registry hack or simply unregistered the DLL that is handling the preview functionality. I would suggest the latter as the registry hack is not 100% functional. So here it goes, select run from the startup menu and enter:

regsvr /u shmedia.dll

To remove the video preview functionality. And to restore it again just enter:

regsvr shmedia.dll

Curse that cursor

2005-04-19T10:02:00.000+02:00

There is a lot of confusion regarding the use of cursors in WinForms. I came across this a while ago with an application that wraps standard WinForms in its on layout. The problem was that the WaitCursor could not be correctly set at all times. Investigating the problem I found that people use one of two models for changeing the cursor, with this.Cursor = Cursors.WaitCursor; and with this.Cursor.Current = Cursors.WaitCursor;. So on to the question What is the correct way and whats the difference?

The correct way (at least most of the time) is actually the simple:

this.Cursor.Current = Cursors.WaitCursor;
// Do your stuff

But hey, the cursor isn't restored to default, but it works anyway? Well, in this case the current cursor is set to WaitCursor (the current mouse cursor). After the event code has been executed it is reset to the default cursor for the control where the mouse happens to be located. This is done automatically. And in most cases this is exactly what you want to do.

But the other way then? The other way, which happens to be what you see the most is the following:

this.Cursor = Cursors.WaitCursor;
// Do your stuff
this.Cursor = Cursors.Default;

Here you actually sets the default cursor for the WinForm, and after you've done your stuff you reset it yourself. Seems simple here but in a complicated if/then/else with with multiple exit points you end up with an hour glass when you wanted the arrow. The problem I found was that in the original WinForm it worked OK but when i moved out to the wrapper application it became an arrow.

The problem you may encounter is when you handle events during your eventcode with DoEvents(). Here you should save your current cursor, set the wait on the TopMost WinForm, DoEvents and the restore the cursor. And make sure to restore it at the right position...