My current job functions keep me mostly in the data realm, there is not much coding I do that is not inside the nice cozy little Query Analyzer windows...but I would be lying to say I have not been intrigued by the new Visual Studio betas. The main reason I had not pu into effort into installing the newer Visual Studio was the experiences of a friend of mine (read Jayson's install experiences here) and all the troubles he ran into. Well, I figured since I had just flattened the machine (again) that there was no better time to try the install than on a fresh OS instance.
I came into this task knowing that when I installed SQL 2005, well, I ended up having to flatten the box then to get it to install, *but*, that was mostly my fault. I had SQL 2000 installed as the default instance and wanted SQL 2005 to be the default instance. I uninstalled SQL 2000, cleaned out the file directories and the registry keys that I thought were relevant. I had worked in support of an application that used MSDE as its backend and a failed install often meant file and registry cleaning, so I thought I was well prepared for the task at hand. Well, after the cleanup, SQL 2005 informed me it did not yet support upgrading from SQL 2000 and I would need to choose a different instance for the install or install as a virtual server. Well, I don't know much about the new virtual server functions in SQL 2005 so I decided against that option. An hour of registry search and delete later, SQL 2005 was still informing me of its lack of desire to upgrade my non-existant SQL 2000 instance. Out comes the OS disk for a wipe and reinstall.
For those of you who know me well, you (the two or three people that actually read this page) know quite well that flattening my machine is by far a rare occurance. Some even infer that it is some sick obsession of mine. Back to the point though, the install of SQL 2005 went flawlessly on the clean install of Windows 2003 server.
Back to current day, the machine has been flattened again, well, a couple times actually, and I decided with the issues Jayson had with installing Visual Studio 2005 that I would install that immediately and then follow with SQL 2005 and install my normal suite of applications following that (VS 2003, VS 6.0, Office, Map Point, Far Cry...). The Visual Studio install went very smooth, and I have to admit, the fit and polish of the installer was very nice. Microsoft's current direction seems to point to catering to the lowest common denominator, meaning allowing people that would normally use Access or Front Page to feel as though they can design enterprise level application, and the new wizard seems to strongly display this. There were all kinds of pretty splash screens you would normally see in a Windows Millenium setup...not that I ever installed Windows Millenium, mind you *cough*, but the look and feel was very end user centric instead of developer centric.
So, with Visual Studio installed, I pop out the disk and slam in the SQL 2005 disk, punch the setup icon and...well, SQL 2005 tells me it won't install because it is incompatible with Visual Studio 2005...
I think that needs to be repeated, and let me preface it by saying I do truly understand what the words *beta software* imply, but SQL 2005 telling me that it is incompatible with Visual Studio 2005...sigh...
What is your application more likely to need than a database backend? And, for those that say 'bah! install them on two seperate machines', well, um, Reporting Services requires Visual Studio to be installed on the same machine as SQL Server...got you there huh?
Now, those few people that know me would also tell you that I am a die hard advocate of Microsoft, well, at least that I was. I have seen nothing but disappointments from them lately, but this isn't a bash Microsoft post. I want nothing more than to see Microsoft regain what Microsoft was, but once that downward slide starts, just how hard is it to stop?
IE is a leaking siv, and until the threat of the Google Browser showed up, they had no intention of making a new Internet Explorer available until Longhorn's release. Even though they conceded to possible make some updates to IE, last I read, they would only be availble to Windows XP users, and assumably Windows 2003 users, but I must admit, a friend from Microsoft yesterday reminded me that I am one of the few people that use Windows 2003 Server as my desktop OS, so maybe not even there.
And, don't even get me started on Longhorn, between the constant slips, and then the announcement that WinFS would no longer be included...speaking of, let's look at that for a moment. The strengths of the new OS were going to be security, something Microsoft has shown a valient effort in improving upon. That in no way implied they have been successful, but hey! an 'A' for effort. The next thing was going to be Avalon, the new desktop that is rendered in DirectX and has all the wonderful new...well, stuff. From what I have read, this is being scaled back to a shadow of its former self, but, you know what? I don't care, I use the Server OS, so obviously all the 'bubbly, cutesy' addons of Windows XP desktop don't mean alot to me, but I was looking forward to the new desktop management and task switching functionality that was to be piggybacked on Avalon, and it may still be there.
The next improvement was going to WinFS. For those of you that don't remember, this was also promised in NT 5.0, which of course became Windows 2000 Server. Along with that was Windows DNA, which for the most part became COM+ in its scaled down release. I know that Windows DNA still had a meaning for awhile but it was by no stretch of the imagination what it was intended to be. Windows DNA was supposed to allow multiple servers to work together in an intelligent manner to decide where in your network had the resources to run the components your application needed. Server A is running 4 components and the network aware runtime notices that the server is being over utilized, but Server B is humming along in nap mode, it shuffles components around to balance the workload, and the current location of the components are kept in Active Directory, so the clients using them can find them dynamically. That is a very simplified example, and many of you will think I am crazy, we don't even have that functionality now. Well, we don't have WinFS now either, nor will we in 2005...besides, if you want to read the amazing things that would have been if they could have been, see if you can find an old Wrox boot called 'Designing Distributed Applications with XML, ASP, IE5, LDAP and MSMQ', the ISBN number is 1-861002-27-0. It is awesome reading. Our tasks as Enterprise Level Developers would be much different if the things in that book became fact, but I am sure we can just wait for Windows Server 2007...
Now the last thing I want to cover as one of the improvements is Indigo, the newer web services implementation. Now most of what I have to comment on here is conjecture, since there hasn't really been anything written on the effects of the lack of WinFS on Indigo, but, considering what WinFS is, the ability to find a file or resource, whether locally, on the intranet, or possible even on the internet, I can not see Indigo being what it was originally planned to be. To duplicate the discovery code that would have been in WinFS anyhow into Indigo would have been rediculous, so I am going to have to assume that Indigo would have piggybacked on WinFS, so we are looking at another technology that will be greatly affected, if not crippled by the removal of WinFS.
If I remember correctly, I gave my oh so humble opinion earlier that my impression of SQL 2005 was nothing more than a patched up SQL 2000 with a new, prettier, but slower interface. Well, maybe not slower interface, but definantly laggier interface. Maybe that just a side effect of the jitting to a beta version of the framework, but I am not holding my breath. And, I am going to predict that Longhorn will be very similar. It will end up being Windows XP SR-2ish, with a prettier but laggier desktop. I hope I am wrong, but I am not holding my breath on this one either, especially since it looks like I would have to hold it until at least 2006...