In my years on this earth I have used many, many MP3 players on my desktop, both Windows, Mac and Linux. My current favorite is Foobar which at first glance seems like the weakest of all the MP3 players (especially compared to Winamp or iTunes) but with a little bit of customization I think it stands head and shoulders above the rest.

Here is a list of reasons why I like Foobar more then any other MP3 player.

1.) Customization. You can customize the heck out of this thing, how titles show up, color, font, etc. If you want to make the play list resemble iTunes or Winamp its more then possible to do that.

2.) Player Window. The player window is totally resizeable. Unlike Winamp which doesn’t resize (except the playlist) or iTunes which generally eats up too much real estate for my taste the Foobar player can be resized in any direction as big or small as you like.

3.) Speed. I have a lot of MP3 files and Foobar is by far and away the fastest of the players I have used in terms of loading then entire library, searching and not stumbling or bumbling during playback. I have had complaints with performance with both iTunes and Winamp.

4.) Advanced Sorting. Foobar has a lot of sorting options, you can even build your own sorting format using If statements which I’ve done. Some of my MP3 files are missing tag data so I can build a sort that uses either the tag name or if its missing use the filename. Just about everything is properly sorted. Windows MCE does a similarly good job but is much to slow to be considered anywhere near my favorite player.

I can modify these settings to album and track number if I want to listen to complete albums, but for the purposes of finding specific songs I like to sort by title after artist.

5.) Codecs and Plugins. Foobar supports every codec I’ve ever wanted it too (like ogg, flac, aac, etc). It is plugable like Winamp so you can add components if one is missing and there are plenty of popular plugins available like Last.fm and audioscrobbler.

If you spend about ten minutes working at it you can setup Foobar to be your ideal desktop MP3 player. You can use the ideal sorting, look and feel, I’ve even programmed all my shortcut keys to resemble Winamp one’s so I wouldn’t have to learn new shortcuts. The performance, smaller footprint, better sorting are all benefits. There are a ton of other customization options for this program and while not perfect its definitely at this time my favorite. I would add that the first time I ran Foobar I really didn’t like it but after I spent a little time on customization it quickly became my favorite.

I first noticed this when working with some images from Microsoft. A 7GB Windows OS image is packed into five 716MB files (total of 3.5GB). The compact is pretty nice but probably comparable to what you can do with Winzip.

Where WinRAR separates itself is how you create the .exe files. Its a simple enough process using the WinRAR wizard. The big feature here is of course being able to specify a max file size which WinRAR uses to chop one big file into multiple more reasonably sized one’s.

Its possible to select from a pre-existing set of options or to specify in bytes your a specific file size.

After the RAR file is created there’s a couple other things you can do. For one if you right click on the archive you can see some of the compression statistics.

You can also right click on the .exe part of the archives and select Open in WinRAR. Once the WinRAR IDE is open go to Tools –> Convert to SFX. When the SFX properties window opens click on the Advanced settings button.

And now you can really trick out the self-extracting executable file. You have all kinds of options, you can create shortcuts, program file entries, use custom images and icons, even insert a EULA or license agreement that will pop-up when a user first runs the executable.

When I create an archive I usually like to use 100MB file parts which I believe is a little more manageable for the images I work with (creates 4-8 parts). Then for storage purposes I zip them up in Winzip which using the Slowest most Compressed option which regains a little more space and makes it easy to copy to other locations.

When creating a re-distributable executable its a nice touch to use custom icons, images, titles and messages. These might seem like little cosmetic things but its a nice touch and makes the image look more professional.

Usually I would run Linux on VMWare, as most people would since Microsoft and Linux have a history of not mixing well. This is not just a rumor as Linux and VPC do not mix very well together either.

My first experience of trying to run Fedora 7 on Virtual PC 2007 was very bad. The installer will not work in Graphic mode so it is necessary to run in Text mode which is fine. Due to an apparent driver problem with the VPC when Fedora boots up the first time the video is corrupt and the interface impossible to use. There are apparently other issues with the mouse and PC clock that if you get past the video problem will plague your install.

Two days and lots of coffee later I’ve become an expert at doing Fedora 7 installs on VPC. Doing eight or nine Google searches finally brought me to an excellent article on what needs to be done to get the install to work.

http://theidealcopy.blogspot.com/2007/09/running-fedora-7-on-virtual-pc-2007.html

I’ve developed my own slightly different approach that I found was just a little quicker when doing the VPC install. Here are Pat’s steps to installing RedHat Fedora Core 7 on Microsoft’s Virtual PC with super excellent results.

1.) Install Fedora using Text mode.

2.) Once the install completes and the VPC reboots for the first time press any key to enter additional boot options.

3.) When the Grub menu comes up select the Fedora install image and use the ‘a’ key to insert additional kernel arguments.

4.) After the work ‘quiet’ in the kernel arguments line enter the text ‘vga=791′. This will solve the resolution problem at boot time and allow us to fix the other items. This also means that it is not necessary to boot up into Recovery mode to make changes to the system.

5.) Press enter and the system will boot up with a visible, non corrupted screen.

6.) Log into the system as root and then modify the xorg.conf file. The command will look something like this.

#] vi /etc/X11/xorg.conf

Pree the letter ‘i’ to get your editor into “insert mode” and follow change both 24 values to 16 under the “Screen” section. Also comment out the s3 driver line and replace it with the fbdev. It should look like this in the “Device” section.

Use the following commands to save the file in vi.

Esc
shift + ;
wq
Enter

(The wq are the write and quit commands).

7.) Time to modify the boot arguments. Use vi again with these commands. 

#] vi /boot/grub/grub.conf 

Rewrite the kernel line to add the “vga=791 i8042.noloop psmouse.proto=imps clock=pit”. The best way to do this is copy it from Internet Explorer on Windows (if you host OS is Windows) and use the Edit -> Paste menu in your Virtual PC toolbar which will type the text into your vi window. Make sure you setup your vi so that it is in insert mode and th keyboard prompt is right after the word quiet. 

 After you save the changes then you are done fixing the mouse, PC clock and VGA screen problems with your Virtual PC. If you also want to boot up into the Graphical Display instead of the command line then perform the last step.

8.) Edit the inittab file using this command:

#] vi /etc/inittab

Modify the line that reads “id:3:initdefault:” by changing the 3 to a 5.

Now you’re done and ready to use your new Linux Virtual PC. I recommend at this time stopping the Virtual PC machine making a copy of the .vhd and .vmc files. This way you have a reusable template in case you need to start over or want to create a new VPC running Fedora.

I am not sure if this problem persists on other Linux distros, I have only seen it described for Fedora.

 1.) Change your network usersname and passwords. The account you use to login to your office network needs to be changed first. I lost a Mac so I did not have an A/D account on the systems but I did have VPN and Email connections to the office.

2.) Contact the IT department about the loss. They can think of areas of concern that you may have forgotten and look for repeated authorization failures in your username.

3.) Change all your personal account information. Email, credit cards, bank accounts, web sites, airlines, anything web based where a username and password is used. Access to these accounts maybe available through web browser cache and cookies, changing the password should prevent this type of access.

4.) Monitor bank and credit card statements for unknown transactions. If a thief is able to steal your identity the first sign will be activity on bank or credit accounts.

5.) If the laptop was stolen you want to request a police report. This may have to be requested in writing or through fax but can take a couple days. The police report is handy to have especially when working with the insurance agency so you have the details of the crime.

6.) Contact your insurance agent. Make sure to read your policy! The claims adjusters can make mistakes so make sure to read through your policy. The adjuster will have a copy of your policy as well and you can step through it together on the phone if there are any questions on item payouts.

7.) Locate the receipt for the laptop. If you have insurance on the laptop or home owners insurance you will likely need to send a receipt to the insurance company.

8.) Search your Email for sensitive data that may have been stolen. I recommend to everyone using Exchange Email server to archive off Email older then a year on a desktop (less portable, less likely to be lost or stolen). This way if a laptop does get lost there is only a year’s worth of Email accessible on it (I keep all my emails).

9.) Check eBay and Craigslist for your laptop. It’s a long shot but it happens all the time where a stolen laptop ends up online for sale. It’s worth it to do a search on used laptops with similar properties. Sites like eBary show the location of the seller so you can search in the area you lost the machine in. Who knows maybe you’ll get lucky.

10.) Start preparing for recovery. Get copies of backup data, software installation disks and license keys. Burn as much as possible to a DVD disk to make it quick and easy to rebuild and restore your laptop once its replaced.

 Hopefully you do have backups of your important files. If you do not have a backup process in place then consider using one of the many online backup services like www.idrive.com and www.mozy.com that are free and easy to use.

“I recently promoted a change from my Workspace branch to our Mainline branch. I did a rollback to an earlier version in my Workspace branch but when I try to do a rebase it tells me there is nothing to Rebase”

Let me explain why that it is.

Surround SCM’s branching engine only applies new changes to your branches when a promote and rebase are performed. Surround uses common ancestor detection and some internal flagging mechanisms that allows the system to always know the difference between new changes to a branch and pre-existing changes to a branch.

When Surround performs a rollback (either a file rollback or promote and rebase rollback) the previous version history is not discarded, so when a promote and rebase are performed after a rollback Surround still detects the earlier changes and will not apply the same set of changes a second time.

Here is an example I came up with.

A file exists called foo.h that is identical between the Mainline and a Workspace branch. The original contents of the file will be called A and so looking at the two branches I can say the file contents are A, A (Mainline, Workspace).

The user checks in a new change (B) to the Workspace instance of the file. Now we can say, between the Mainline and Workspace the file contents are A, A + B.

The files are now different between the Mainline and Workspace but when the User tries to perform a rebase Surround will show no new changes, because the contents known as A already exist in both locations.

The new change (B) is promoted to the Mainline. Now the file is identical again in both locations A +B, A + B.

User performs a rollback in the Workspace and now the file contents are A + B, A + B - B.

It appears that a newer file exists in the Mainline since a Diff shows just the contents of A in the Workspace and A + B in the Mainline. When a rebase is performed Surround will still say no new changes. This is because both changes A and B are still present in the Workspace branch. Change B cannot be reapplied a second time.

This is what Surround see’s; A + B, A + B - B and of course changes A and B are already present in the Workspace foo.h file.

If the user wants to get this file back to the pre-rollback contents then ‘another‘ rollback must be performed. The file contents now look like A + B, A + B - B + B resulting in no differences between the Mainline and Workspace. (The file history will show 1, 2, 3, 4 for the add, checkin, rollback to version 1, rollback to version 2).

To Summarize;

• Surround does not discard any history information, even during rollbacks.
• Surround does not allow changes to be re-merged more than once (I call this merge memory).
• As a result, a rollback will not result in pre-existing changes being re-applied during promote and rebase.

The merge memory feature is one of the outstanding benefits of the Surround SCM system, as it prevents any changes you make from being merged multiple times. Compare this to a system that does not keep track of these merge points and would allow the accidental reapplication of an old change which may have a negative result, like undoing a bug fix.

Recently, the captain of this ship called Seapine, gave me a MacBook Pro to use for work-related activities. I took this as a hint to move to a more Macintosh lifestlye and so without a second thought I have been using a Mac every day now.

Moving from a full time Windows and part time Mac/Unix person to a full time Mac to a part time Windows was a little weird at first. I already have lots of experience on Mac and with UNIX so it wasn’t a technical hurdle, but it is a lot like buying a new pair of hockey skates and taking a few days to break them in so they feel right.

Since MS Office runs on Mac a lot of the everyday business work is unchanged. I would add that Excel on the Mac is a little odd to me, but everything else like Word and PowerPoint is pretty much the same. Entourage is the Email client that replaces Outlook. Its pretty much the same thing although I actually like some Entourage features better but some of the configuration aspects are difficult to find our nonexistant. Overal its not a stretch from Outlook.

The first week I was setting up the MacBook Pro I installed Adium for IM (the best Multi-Protocol IM but no iSight support), GoogleDesktop, VMWare Fusion, Eclipse, Xcode, FireFox and Mono.

So far I happy with all the Mac software I use that replaces the Windows software I previously used except of iTunes. I really do not like iTunes, it takes up too much desktop space and it has too many features I don’t want but cant disable (like automatic album downloads). There is no good alternative to iTunes though I have tried about a dozen apps to find something more like Winamp (Cog, VLC, Whamb, Songbird, Play, to name a few). Cog was closest to what I was looking for but its too unfinished of a product and unstable with large number of files.

I also got a little into .NET and Mono which I will talk about in my next post!

Overall I am pretty happy with the switch. I am still not a huge Apple fan (mostly because of cost) but VMWare Fusion looks real good and it allows me to run whatever Windows only apps I might ned (like Visual Studio). I also tried CrossOver which is a WINE port for Mac and I really wouldn’t recommend it. Using some of the more advanced features accessed through UNIX (like using X11 apps) can lead to more tricky Unix issues, but otherwise the Mac is a tight system with a lot of good apps you can download from Apple.

It is true what they say about the Mac, about how things ‘just work’. Anyone who has ever tried to setup Bluetooth on Windows probably knows what I am talking about. Mac features like iSight, built in Mic and speakers, Bluetooth, IR, everything really does just work without much effort. Using iChat I was able to do Video Chat in about five minutes from the first run.

I’ve been in Denver for about two weeks know and have just been spending most of this time getting settled into the new house and making sure all the home office equipment is working. The IT group set me up with a fancy Cisco 870 Router which, when I am plugged into, keeps me constantly connected to the Seapine VPN. Without having to run any special software I am on the network and my Cisco VoIP phone, Outlook and network services are all running.

For Customers and Co-Workers this means I can still be reached using the same phone number and it’s almost like I am still in the office (except mountains are outside of my window instead of the Kings Island Amusement Park).

A nice aspect of living in Denver (besides mountains) is that many cities in the area, Boulder, Colorado Springs have growing IT industries. Many Seapine customers come from these areas and this will enable us to bring an on-site presence to a lot of new and existing customers. It will also enable easier travel to the west coast and California where a lot of Seapine’s business is done. Being 2 hours closer to the west coast will make it easier to serve that market as well.

Now that I am moved in and the home office is up and running I guess its back to work!

An interesting thing I discovered when writing this program is there does not seem to be a very good way to parse HTML. I asked one of the experienced developers here who suggested trying to treat the HTML response like XML. That didn’t seem to work since HTML is not structured like XML, its a little too loosey goosey.
I did some searching around the web and found a guy who had written his own HTML parsing classes. That looked interesting, but seemed way to complex for what I wanted to do. Int he end it looked to me like if someone was looking to parse HTML in a very structured and proper way then creating your own classes is really the best solution.

I ended up converting the HTTP response into a string then tearing the string apart until I had what I wanted. A trick I learned was to use the Replace() function to switch out some semi unique characters in the response like a and replace it with a “^” which then makes it easy to Split() the string into an array. I can then clean up the array strings easily. I used a lot of .IndexOf() and Remove() string functions as well. Starting with the HTML from an entire page I would remove everything before what I wanted, everything after what I wanted then clean up whats in between or Split() into an array if needed.

That approach works, but it was very messy/ugly/kludgey. I see how XHTML might make this whole act easier. My first attempt was to use Regex, and after I developed I migraine I decided to just start ripping apart the strings. This is probably a more lowbrow approach but more appealing to my senses then regex.

You can see the TestTrack Report Publishing script here.

TestTrack_Automated_Report_Publishing

There are some very good tips I read a long time ago when Seapine was first getting into the automated testing market, and these tips are still very true today.

http://www.jeffgainer.com/testmyth.html

Jeff’s short article is designed to prepare users for the reality of automated testing and to not get too caught up with the excitement that automated testing brings with simple rules like not to get to complex with automated testing, make a real investment into the process (meaning time) and remember that automated testing does not eliminate manual testing .

There is an article I also read at the same time from Bret Pettichord which lays out the sever steps to automated test success. Its a very good read and I remember this article whenever I implement or present our solution to organization. Just like with manual testing the better organized and the more you plan and outline your process the more success you will have with your testing solution. And the article by Bret provides a nice level of detail to describe how that might work.

http://www.io.com/~wazmo/papers/seven_steps.html

There’s always a sense of real relief that washes over a QA testers face when they find out they can automate the 10,000 data entry tests they had been performing manually. Constant repetitive tasks are mind numbing and turning those over to the machine is always a welcomed change. I think data driven automated testing is one of the most beneficial features of automated test tools, but there is a lot you can accomplish with tool like QAWP. The more scripts that are written the larger your regression suite becomes which means as your software applications increase in size and complexity your QA regression suite grows with it. When done successfully QA testers can focus on manual testing of new features or complex tests and the machine can run the regression and data driven tests prior to each build. Some groups even use automated tests to regression or smoke test application builds on a monthly, weekly or sometimes daily basis (unit testing is better for continuous integration testing as automated tests usually take longer to run, though automating smoke tests could be continuous depending on build frequency).

If you haven’t already head over to http://labs.seapine.com and checkout the new QAWP beta. Its a big improvement over the last version of QA Wizard where the biggest noticeable difference is the VB scripting language used for scripts themselves.

Getting back to the Jolt Awards. I was able to attend the ceremony this year which was hosted by Craig from craigslist. Part of the ceremony was done using Second Life. Seapine won a Jolt for TestTrack in the bug tracking category. We’ll be working hard this year to see to it that Surround SCM and QAWP make it to the finals next year!