markmanoukian.com

I wrote some code to upload documents to NetDocuments via the NetDocuments’ SOAP APIs. For no apparent reason two of the API calls would not work. Fortunately, I had a baseline by which I could compare my failing code with a piece of working code — the NetDocuments E-mail Management System (EMS) add-in. EMS makes the exact same SOAP calls that I was using and EMS works.

NetDocuments tech support recommended that I use Fiddler2 to capture traffic to better diagnose the problem. Comparing the successful SOAP calls from EMS and the failing SOAP calls in my code showed that the problem was not in the body of the HTTP request, but rather the SOAP binding, which is the header. NetDocuments required a different user-agent on the failing SOAP calls.

I could not have done it without Fiddler2. Fiddler2 is here: http://www.fiddler2.com/fiddler2/

Fiddler2 captures http traffic

I’m playing with themes today. I like a clean look, which to me means a white background. I want a san serif font for headings and a serif font for the body. In the sidebar I want links to recent entries, and perhaps a list of months in which I posted. I need to do more with tags\categories. I was using Atahualpa until today. Water looks pretty good, but I think I’m going to run with Twenty Ten for a while.

Browser Master prompts you to choose a browser when opening a hyperlink. If, for example, I have a shortcut on my desktop to http://www.osu.edu/ then when I double-click on it I will be prompted to choose the browser in which I want to open the link.

My favorite example of how useful this can be is that I rarely use IE as my default browser, but I often use LiveMeeting for web conferences which works best in IE. It used to be that when I was reminded or received a LiveMeeting link I’d click on it and then spend a couple of minutes backing out of it, and copying and pasting the link into IE. Now, I simply click the link and choose IE instead of Chrome/Firefox on the fly.

The look, feel and graphics are super cool. Upon double-clicking a row of four oversized icons for IE, Chrome, Firefox and Safari fade in. I can choose whichever browser I want and the URL will open in the selected browser. If I do not choose one then after two seconds it will default to Chrome.

If I want to remove a particular browser from the prompt without having to uninstall it from my computer, then I can do that by simply editing the config file for Browser Master. If I want a timeout of more than two second then I can do that too.

If you want to try it then use the 1.1 beta. The 1.1 beta works better than the 1.0 stable version. The 1.1 beta runs reliably, but the installation requires manual tweaks.

I ran into a few obstacles all of which I was able to resolve. I tried Browser Master on Windows XP, Windows Vista and Windows 7. It did not work on any of those systems out-of-the-box because it was not properly configured. The configuration tasks can be broken down into two categories: 1) necessary edits to Browser Master’s configuration file, and 2) necessary edits to the Windows registry.

Browser Master’s configuration file is an XML file. The config file will be placed whereever — i.e. it depends on whichever OS you are running. You can determine the location by running “BrowserMaster.exe -e”. Typically that will open the browsermaster.xml file in IE, which is completely useless for editing, but this will also show you the full path to the XML file so that you can now open it in something such as SC1 or Notepad++.

Once you open the config file in your XML editor then do the following:

1) Delete the browser configuration for browsers that you are not using. Why? Because BrowserMaster tends to throw errors for things are specified in the config file aren’t actually where specified. And, don’t worry; if you happen to install one of those browsers at a later date then you can easily restore the config for that browser.

2) Check the executable paths of all of the browsers. Why? Because BrowserMaster tends to throw errors for things are specified in the config file aren’t actually where specified. Remember, if you are using Vista or Windows 7 then the path to a particular browser might be “c:\program files (x86)\…” and not simply “c:\program files\…”. I had to add that “(x86)” to several lines of the XML. The other thing to remember is that some browsers — e.g. Google Chrome — may be installed entirely in the user profile, in which case the executable path might be “c:\users\%user profile%\…” and not “c:\program files…”

3) Check the paths to the icons. Why? Because BrowserMaster tends to throw errors for things are specified in the config file aren’t actually where specified. In my experience the paths for the icons were folders in the user profile and had to be edited to reflect the correct locations in the BrowserMaster program directory.

Once you’ve got the config file sorted out you can begin testing. Create a shortcut on your desktop to “…\browsermaster.exe http://www.google.com”. Double-click the shortcut. If you are prompted to choose a browser then congratulations. If you are not prompted to choose a browser then double-check the config file — BrowserMaster.xml. If you want to keep it simple then delete the configuration data for all but two browsers. If you want to keep a backup copy of the config file so that you can restore config data easily then you can certainly do that. If you have deleted some of that data and you want to recreate a new configuration file then you can do that by running “…BrowserMaster.exe -d”. (Note: if you have an existing configuration file then running that command will either overwrite the existing file, in which case you’ll lose previous edits, or it will not do anything, in which case you will not recreate those lines of xml that you previously deleted.)

Regarding registry edits Browser Master threw errors on installation on more than one occasion along the lines of “Can’t delete registry key because it doesn’t exist.” leaving the installation incomplete. Honestly, I don’t care because I’m not a fan of the default method by which Browser Master configures itself in the registry anyway. I prefer to see it installed like other browsers such that it is listed as a “browser” from which I can choose in the “Default Programs” dialog. To accomplish this Browser Master has to be included in a) the list of registered applications, b) the list of web browsers as shown “StartMenuInternet” in hkey_local_machine\software\clients, and c) it needs to be registered in hkey_classes_root. See the attached reg files for an example.

Notes on the attached reg files: a) if you see references to Chrome in these files that is because I chose to default Browser Master’s icons to Chrome icons because I like them better, b) note instances wherein the path to the user’s profile is included which you can find & replace by searching for “[User Profile]” as a precursor to merging these reg files, and c) you’ll have to change the file extensions from txt to reg:

a-BrowserMasterRegisteredApplications
b-BrowserMasterStartMenuInternet
c-BrowserMasterHTML

Browser Master is here: http://browsermaster.codeplex.com/

Special thanks to jquintus and jrall for this brilliant piece of software.

Thx

A colleague reminds me that AutoIT, another open source scripting solution, offers some of the same features and functions as Sikuli. Indeed, it is possible that it does everything Sikuli does and more. I haven’t taken a look at AutoIT yet, but there it is.

AutoIT is here: http://www.autoitscript.com/

Found on Lifehacker: A little app called Sikuli. (http://lifehacker.com/5459673/sikuli-automates-almost-anything-with-screenshot-ease)

Sikuli allows me to write a script that automates GUI interaction — e.g. button clicks, box checking, typing, etc. — that would otherwise have to be performed manually due to the absence of a command line interface. Many apps have command line interfaces (CLI), but every once-in-a-while we come across an app that doesn’t.

Lack of a CLI can be a real nuisance on some occasions. In my own experience I have an app that needs to run at all times on a desktop in our server room. It would be great if the app would simply start whenever the server restarts, whether the restart is due to windows patching, maintenance on the hypervisor, whatever. Historically, we are able to configure the app to launch, but it doesn’t really do anything until a particular tab is selected and the start button is clicked. With Sikuli we are able to automate the selection of the tab, and the clicking of the start button.

It’s in beta, but it works quite well.

It runs in Java, you need Java.

Because it interacts with what is showing on the screen, be aware that it cannot see controls or other objects that are behind other windows. It cannot double-click a desktop icon if, for example, desktop icons are hidden.

I saw many questions to the issue of how to run a Sikuli executable as an executable, because it isn’t immediately apparent and there were inconsistent answers. The best thing to do is to use the export as executable command, which exports a Sikuli script with a “skl” file extension, and then associate that extension with the Sikuli*.bat file that is installed when you run the Sikuli setup.

It works.

Sikuli is here: http://groups.csail.mit.edu/uid/sikuli/

Here is something that I submitted to Technolawyer (www.technolawyer.com) in late 2007, that was published in April 2008…

I enjoyed reading your article “How I Use Linux in My Small Practice” and heartily endorse your conclusions. Linux and the larger open source software movement have a tremendous potential to improve the software experience, from development, to acquisition, to implementation, to usage.

The general public often equates Linux with open source.

Linux is perhaps a cornerstone of the open source movement, but it is still only a subset. Many-to-most computer users are unaware that they are using open source software, and specifically Linux, on a daily basis.

As Mr. Hubbard observes, much of the Internet is powered by open source software. The Linux\Apache combination of OS and Web server is the most popular Web hosting platform. BBC, Wikipedia, Travelocity, Motorola and BlackBerry all run on this platform. The system by which domains — i.e.

www.microsoft.com in a Web address or “Microsoft.com” in billg@microsoft.com — are resolved into IP addresses is built around open source software. In essence, if one uses the Internet then one is using open source software.

We find open source software in many other places as well.

Linux is embedded in many devices that we use today — e.g.

SANs by FalconStor, the Linksys WRTG54 router or cell phones by Motorola and Samsung. The leading CRM application in the legal industry, Interaction, incorporates the Tomcat Java Servlet. If you use Interaction, you are using open source software. Services such as Postini and NetDocuments incorporate open source software. If your network administrator is aware of open source, you might be connected to the Internet through an open source router — e.g. IPCop — or maybe the monitoring systems in use could be open source — e.g. Nagios, Cacti, SmokePing, etc.

There are many, many open source desktop applications for *Windows*. VLC is a wonderful media player, that works where other media players sometimes fail; it includes codecs built-in to the application so you don’t have install Quicktime or RealPlayer to watch a particular video. Use Inkscape to edit a scaled vector graphic instead of Adobe Illustrator. Don’t know what a scaled vector graphic is? Go to the Inkscape online documentation; it explains in layman’s terms what a scaled vector graphic is, how it is different from a bitmap, and how to use Inkscape. Try GIMPShop instead of Adobe Photoshop. Need to produce a PDF; how about PDF Creator?

Firefox and OpenOffice are a couple of flagship open source applications that will run on Windows. What is special about these pieces of software? Well, the open source community has been accused of stealing from the commercial community, but Firefox turns this argument on its head. The first browser, Mosaic, came from the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign. Firefox was the first Web browser to incorporate tabbed-browsing. Firefox is *easily* extensible; all the little bells & whistles that can be bolted on to Firefox are available in such abundance because the programming is performed in a text editor, whereas in Internet Explorer extensions have to be compiled into DLLs.

This means nothing to the layman other than understanding that Firefox is literally open to customization by anyone, and therefore enjoys extraordinary global support.

OpenOffice truly works quite well, and is perhaps more like Microsoft Office, historically, than Office 2007. Anyone that can use Microsoft Office can use OpenOffice. Many of us would argue that OpenOffice is better, that it includes the features and functions that are important to us. While you can now download an add-on from Microsoft to export from Word to PDF natively, this function has existed in OpenOffice for years.

OpenOffice also complies with Open Document Format (ODF), an ISO standardized document format that better allows us to exchange documents and mitigates conversion incompatibilities (or the need to convert, all). In contrast, the Microsoft Word format is not only not a certified document standard, it is not even standard across the versions of Word. Subsequent versions of Microsoft Word not only add new features and functions to the format, but modify the existing features and functions. If, for example, you used CompareRite 2000 to redline a document created with Word 2000 then it will work. If you open that same document in Word 2003, make no changes, and save it to Word 2003 format, then you can no longer redline it in CompareRite 2000. The document has actually been converted to a significantly different format, not much different than if you had opened and saved it in WordPerfect. The Microsoft Word format is the antithesis of a standard. Imagine a world where we no longer have to convert documents when sharing them. Setting and adhering to standards is one of the larger objectives of open source software. You’ll find such compliance time and again with open source software.

We use these pieces of software because they are excellent and they eliminate obstacles. Price is only one obstacle. If you are like me, you might find yourself needing to work while on vacation. I once found myself in a situation where I needed to produce master copies of a CD with artwork for distribution at an upcoming conference. We needed 2000 copies of the CD. I downloaded, installed and used GIMP for general image editing, InkScape for SVG editing, PDF Creator to produce various PDFs and InfraRecorder to burn the masters. While many critics perceive that open source software survives on lack of price, the truth is that it survives on merit.

I recommend that novices start with the Ubuntu 7.10 “Gutsy Gibbon” CD. Ubuntu has done something clever in that their CD performs three functions: 1) if you slip it into your PC while running Windows then it will autolaunch and prompt you with a menu of handy opensource applications that you can try under Windows; 2) if you boot off the CD then it will run Ubuntu off the CD like Knoppix; 3) if you boot off the CD then you can also install Ubuntu, side-by-side with Windows (assuming that you are running Windows XP or older; dual booting with Vista is technically possible, but not easy).

From http://www.energystar.gov/index.cfm?c=power_mgt.pr_power_mgt_ez_gpo …

EZ GPO is a free tool for network administrators who run Windows Active Directory and manage client workstations using Group Policy Objects (GPOs). EZ GPO provides Group Policy Objects for centrally configuring power management settings, as they are not native in Windows 2000 or XP. However, power management can be managed through Group Policy on Windows Vista machines. (For this reason, the EZ GPO tool does not work on Vista PCs.)