Sorry Joe. You tried. Hard. But I am not convinced Citrix Streaming is “the” best solution out there and I will explain why. Note that I will not even get into the technical limitations (i.e. virtualizing services). Does it mean I do not like it or that it does not work? Nope. Certain things I like and it does work. But…
First of all, some background information here. I have been doing application virtualization for several years now, way before Microsoft acquired Softricity and I was directly involved with the first large scale deployment of such thing within the Canadian Government and later on, bringing such technology to several other places in Canada. And as most of my techie friends (the Microsoft MVPs, the Citrix CTPs, etc) as we are early adopters of anything (yes, at least 10 people I know reformatted their PCs back in the 90s to load Microsoft Bob) this means we usually have to do everything and in this case that meant installing all the backend infrastructure, clients AND actually creating the packages for the virtualized applications. Add to that I used pretty much all of them over the years (SoftGrid/App-V, Thinstall, Altiris SVS, VirtualShite, etc), Citrix being the exception here.
After using the latest and greatest offering from Citrix, I am happy to see it works BUT in a nutshell, it is one of the biggest PAIN IN THE ARSE application I have ever used. Why?
1. Creating the application packages. In many places, once I leave, the application people will be the ones responsible for creating these as they were the ones packaging regular applications using products like Wise Installer, InstallShield and so on. So they are used to see certain things and they expect a certain behavior on all these apps. They see the ‘packager’ on any virtualization solution (i.e. Citrix Profiler, Microsoft App-V Sequencer, etc) as what the name implies, a packager. In all these tools when I install application A to C:AppA, I can open My Computer and under the C: drive I will see the AppA folder. Not with the Citrix Profiler. I understand why not, the logic behind that decision and so on. And I still think this is a major PITA. If all the tools these guys are used to, work in a certain way, all the competitors one’s work in a certain way, WHY REINVENT THE WHEEL? No idea.
2. Package deployment. As of today once you finish ‘profiling’ the app, the Citrix Profiler asks you for the share where you want to save it. Once you did it, anyone with access to that share will be able to launch the app, unless you then assign NTFS/Sharing permissions to the folder where you saved the app. Once that is done you need to use another tool to ‘publish’ that app to the users. So you need to select THE SAME GROUPS/USERS AGAIN (the ones you assigned NTFS/Sharing permissions). Why is this bad? Well in many (and I mean MANY) companies groups may have names like ‘Finance – Applications – Microsoft Project’, ‘Finance – Applications – Microsoft Publisher’, etc. So it would be very easy for the person assigning the permissions to the folder to select the ‘Finance – Applications – Microsoft Project’ group and later, when doing the publishing, selecting ‘Finance – Applications – Microsoft Publisher’ as the names may look very similar. What will be the end result? Users will not be able to launch the app. How to fix this? Very simple. The Citrix Profiler could ask if the person saving would like the tool to try assigning permissions to the folder and would allow that person to select the group they want to allow access to the app. Once selected, the tool could set the permissions and store that information on the .rad file so when the application is published, it would already know to which group it must be published and would not ask for that again. End result? The app works for the right group AND you cut down one step in the whole process. The more you can automate, the less human errors may be introduced. Keep that in mind Citrix.
3. I wonder if the guys behind the Citrix Profiler know that environment variables and things like PATH exist for a reason and are actually still used in the year 2009. And yes, you guessed it. The Profiler does NOT seem to capture changes on them. Why? AskJoeNord.com?
4. Why, again, WHY the tool cannot simply disappear, let me do whatever I want to install my app (creating shortcuts by hand, copying files using Explorer, etc) and then capture what I did is another great question. I would say in the same league of ‘Is the Yeti Canadian or American?’ and ‘Where do they keep the Aliens captured in Roswell?’. One of these mystical unanswered questions.
I could keep going on with this list but I will stop here as I am still in Ft. Lauderdale and Joe and his Citrix buddies know where I am staying…
And before Joe leaves a comment saying ‘Oh the tool may not have rights to do that, the person doing it may not have rights, the machine may not be on the domain, bla bla bla’, he knows there are workarounds and solutions for all these cases. So why are these still not there today? Not sure. The only thing I am sure is I will give Joe hell until all this is fixed. 🙂
Seriously I do think this is a typical case where Citrix completely skipped ‘Usability Testing’ and actually ‘Testing’. How come a 5.2 release of anything still ignores PATH being changed? AskJoeNord.com.
On a positive note, the tool, once you get used to it (I mean, once you are forced to accept certain things), works and works pretty well. And with almost no backend requirements what I think is cool.
So Citrix, keep this one thing in mind. The next time, before releasing something like that, give me a shout. I will test it for you.
3,444 total views, 2 views today