So it seems as though I forgot a couple of points in last night’s blog post (First Impressions of Google+).
I should be able to sort my circles
While you can create new circles there doesn’t appear to be any way to sort them. Unfortunately this means that my extreme circles (close friends and fringe) are right next to each other and that’s just not right. However the problem really becomes evident with the menu list of your streams. Only your first custom circle is shown in the list (when you have added more than two custom circles). These are also sorted alphabetically after the predefined circles, which is a little odd, but I can see why it might be the case.
Preliminary conclusion
I think that at least in my case, Google+ is more likely to displace Twitter than Facebook. That said I’m not overly active on Facebook, but I find using twitter is a great way to keep up to date with the people I actually care about and a little industry stuff as well. I think Google+ is well targeted for that particular purpose. Like Twitter though you only know who is actually listening, not how much they care about what you say.
Well, I managed to get on to Google+ today. My first impression is it feels like the love child of Facebook and Twitter. It may look a lot like Facebook when you are viewing your ‘Streams’, but the model feels much more like how Twitter would be if you could tweet to specific groups of people. I won’t bore you with the details of Circles and how they work, instead I’ll point out the things I really like and the missing features that I hope are filled in soon.
The Stream
Well, this is all pretty standard. So all the things from your circles that your circle friends have deemed you worthy to see appear here. What I would like to see here is the ability to exclude feeds from certain circles from appearing in the stream. This would allow your default view to be void of noise but still provide ready access to these people.
Circles and Sets
I have no doubt that this is a feature the guys at Google are keen to push out (the math nerds that they are). Basically what is needed is the ability to define composite circles, that is circles that use standard set operations (union, intersection and subtraction) to define their members. Of course they would have to explain it better than me, but some Venn diagrams should make it clear enough to just about anybody.
Suggestions that are smarter about multiple email addresses
Because the suggestions are based partly on the entire contents of you Gmail address book I’ve found that I get suggestions for people I’ve already added (but have multiple email accounts). One suggestion was even to add myself. Where I’ve told Google that I have multiple addresses I would hope that it could prune some of that for me.
I do like the multiple personality approach that Google+ gives you. I’m glad that even though set functionality isn’t available yet that by using circles you can model some pretty complex social hierarchies. Whether enough people will make the switch remains to be seen. I’m not holding my breath, because unless Google rapidly expands their trial interest may just fizzle. So, hopefully Google can bring wave after wave of improvements while the buzz is still in the air.
Recently I decided to put a bit of Powershell in a reply to a tweet. I was using my fancy shmancy new Windows Phone 7. Well, needless to say I ran into a few problems when I tried to insert the pipe character. I suppose it isn’t too commonly used, but I really wanted it to get my message across (whilst I could have probably used a capital I to get my point across, some fonts won’t do that and it wouldn’t be able to be pasted, so that wouldn’t do).
After holding down what felt like every button to see what special characters they revealed (like the iPhone, the ° symbol is hidden under the ’0′ key) the pipe character still eluded me. However Windows Phone 7 also comes with a special smiley keyboard which has a wide array of smileys to choose from, including the flat :| smiley. Knowing that was the pipe character right there it became easy, simply insert the :| smiley, move the cursor between the colon and the pipe, hit the backspace key and move the cursor back to the end of the line.
It couldn’t be simpler…
Actually, maybe it could. Pipe symbol please!?!
I use Powershell scripts to update test environments and when I can I prefer to use the MSI files that we would ship to a customer rather than hacking together an xcopy deployment. Recently I worked on a script that did the following:
- Check if an MSI already exists in a folder I designated as holding the current installation files (I called it ‘Current’).
- If the MSI exists, execute the uninstall process for the MSI, and remove the MSI from the ‘Current’ folder.
- Copy the new installation file from the build drop folder and put it in the ‘Current’ folder.
- Execute the install process on the MSI in the ‘Current’ folder.
However I soon ran into a problem in that when I called msiexec.exe it would not block the script, so it would try to run multiple instances on Windows Installer and if you’ve had any experience with Windows Installer you know that just doesn’t work (and for good reason).
A quick search on the interwebs revealed that I could simply wait for the msiexec.exe process to finish. Rather than doing some sort of convoluted monitoring of the process inside Powershell I decided to use the the Start-Process commandlet (inspired by Heath Stewart‘s post ‘Waiting for msiexec to Finish‘).
Start-Process is a little bit different from ‘start’ especially in how it passes the parameters (through the -ArgumentList argument/parameter). But fortunately the -Wait parameter was exactly what I was looking for. Here’s the final line:
Start-Process -FilePath msiexec -ArgumentList /i, $installer, /quiet -Wait
This let everything nicely chain together and now deployments are super easy, as they should be.
The Coded UI tests in Visual Studio 2010 are pretty cool. For the last few months I’ve been using the framework to completely automate a sophisticated web application, with lots of drag and drop and crazy design surfaces.
But the test details page has always looked a little verbose. The default Debug trace is full of so much stuff that I find it completely unusable, and what it does have is barely legible anyway. So usually I just collapse the section and move on to my own less verbose logging and the stack trace.
Unfortunately all this information comes at a bit of a price and I have seen machines run into the good ol’ OutOfMemoryException more than once while working with the log. So to eliminate that cause from the list of possible culprits I hit the net and went searching for how to disable the debug trace in Coded UI Tests.
Unfortunately I didn’t find much (other than how to enable, and usually by editing the test agent configuration). I wanted to find a solution that I could add to my solution (damn overloaded words) that would just work no matter what machine I ran the tests on. Fortunately it was really easy. I just added an App.config file with the following:
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<system.diagnostics>
<switches>
<add name="EqtTraceLevel" value="0" />
</switches>
</system.diagnostics>
</configuration>
It did the trick and it is making my life a whole lot easier. Hopefully it helps someone else too. (You could also try other values for the level, but I’m an all or nothing kind of guy).