I've been doing a lot of soul-searching over the past couple of days. This was triggered by some very interesting and useful discussions at the online CodeRage conference this week, along with my own experiences with Vista on Chiara's laptop that I installed this week.
If you've read my other blog entries, you'll know that I wasn't too happy with the initial impressions of Vista. And this week at CodeRage, we have been seeing the upcoming Delphi 2007 product and how it can help make applications work better on Vista.
Even though I haven't been personally impressed with Vista, I have come to the realization that it isn't going away. New computers ship with Vista and more and more of *my customers* will be using Vista, regardless of whether I personally like it or not.
One of the main purposes of CMUD was to bring MUD clients into the modern versions of Windows. To change the architecture so that it would work with new features and allow new and cool stuff to be added in the future.
Now that Vista is here, I can't just ignore it. As a *business*, I am really required to support Vista as much as I can and fulfill the promises I have made for CMUD on Vista.
So, contrary to some expectations, I have decided to embrace Vista, along with the new version of Delphi 2007. This combination will allow me to ensure that CMUD works the best that it can on this new version of Windows.
I don't want to make the same mistake that I made with zMUD. With zMUD, I resisted new versions of Windows and new versions of Delphi. zMUD fell behind the technology curve, and I waited so long that now zMUD doesn't even compile on newer versions of Delphi, and doesn't work well with new versions of Windows. Even the WinXP support in zMUD is pretty minimal. I don't want this to happen with CMUD.
Yes, upgrading to Vista will probably be a pain. Upgrading to Delphi 2007 will require upgrades to several 3rd party component sets, and upgrades to many other tools, such as my profiler, debugger tools, editor enhancements, etc. It's expensive and it will be time consuming. It will probably take me about two weeks of working at night to get a new development system up and running well.
But in the long run, this work and expense should pay off. By actually developing CMUD on a Vista system in Delphi 2007, I will be able to ensure that CMUD works extremely well with Vista and starts to take advantage of some of the newest technology. Rather than developing on WinXP and then just testing it on Vista to see if it runs, I'll be developing on Vista and then using my existing VMware installation to ensure that it still runs well on XP. This is a better way to develop software that I plan to have at least a 10-year future.
Given how long it took for Microsoft to release Vista, I don't think we will see a big new version of Windows for at least five years. During those five years I should be trying my best to make CMUD the best MUD client that I can, and not trying to hide in the sand and complain about Vista.
I am putting together an upgrade plan that will hopefully allow me to install Vista and get a new development partition working without effecting my current development partition. The goal is to have both the WinXP and Vista partitions running on the same machine so that I can continue to work on CMUD on WinXP using Delphi 7 until such time as the Vista+Delpho 2007 system is ready and working. I'll be doing most of this sysadmin work in the evenings so that it has a minimal impact on bug fixing and new features being added to CMUD.
This will delay some of my plans for improving the web site (which I also was going to work on in the evenings). But I think that's actually a good thing, since I want to wait until Delphi for PHP is released to see what it is really capable of doing. With Delphi for PHP I am hoping to make some pretty amazing improvements and changes to this site over the next year. But while I wait for Delphi for PHP to be released, my evening time is probably better spent getting a Vista development environment working.
I'm interested in anyone's comments or suggestions on this, especially if you have experience on putting together a Vista partition on the same system as a WinXP partition. I've read some wierd stuff about the system restore messing up systems that dual-boot XP and Vista. I'm also still trying to decide whether to do a clean Vista install (and reinstall all apps), or make a copy of my XP partition and then do a Vista upgrade (and deal with the resulting problems). I'm leaning towards a clean partition, even though it will be more work in the short-term.
It is important for me to make the final decision on all of this today, since this is the last day for the big discounts on Delphi products for the CodeRage conference.
Firstly - clean install - absolutely no doubt that it's better than an upgrade and you'll have far fewer problems in the long term.
That aside, great news! As one of those awkward Vista users I'd be really interested to see what you will get out of it as a developer. UI considerations aside you'll be able to see what has really changed with the OS aside from just the facelift that most see. You're right, Vista isn't going to go away, nearly every new PC will ship with it soon and in a few years it'll be dominant. Jump on the band wagon early and you'll have the chance to get a much more solid base to improve CMUD in the decade to come.
_________________ CMUD Pro, Windows Vista x64
Core2 Q6600, 4GB RAM, GeForce 8800GT
Because you need it for text... ;)
Joined: 10 Oct 2000 Posts: 320 Location: Springville, UT
Posted: Fri Mar 16, 2007 6:04 pm
One way to have both is to have separate hard drives for the OS versions. I just put together a new system for myself, and found something quite interesting. I left my old XP hard drive intact, and bought a new one for Vista. Both hard drives are SATA. I set the new hard drive as the primary, installed Vista, and now I get a boot menu to let me choose between XP and Vista.
Joined: 25 Sep 2000 Posts: 23377 Location: Colorado, USA
Posted: Fri Mar 16, 2007 6:30 pm
A new disk isn't really an option (money-wise) at this point. I use a mirrored RAID configuration. My motherboard only has 2 SATA connections, both of which are in use by my current disks. Each current SATA disk is only 100GB, with a C: main disk partition of 90GB (all user files, applications, etc), and a F: Windows XP partition of 20GB. I have 60GB free on the C: partition, and 10GB free on the XP partition. So I can pretty easily add another 20-40GB (whatever Vista needs) Vista partition.
Since I only have 2 SATA ports, and they are both used, my only option would be to install 2 bigger disks, and then figure out how to transfer the data from the old RAID array onto the new disks without any spare SATA ports. I also have a 200GB IDE drive where backup images are stored, but it only has about 60GB free right now.
So I think just making a Vista partition is the best option right now. Given all of the other money that I'm going to be spending on various software upgrades, I can't really afford another $300 or so for two new big SATA disks, even if I could figure out how I'd transfer the existing data. Too bad this computer doesn't have more SATA ports.
Joined: 28 Sep 2000 Posts: 1395 Location: United Kingdom
Posted: Fri Mar 16, 2007 7:24 pm
I'm dual booting an XP/Vista system at work on one HD with two partitions without a problem. I've not seen any problems with System Restore so far.
I used PartitionMagic in order to repartition the one full HD partition into two to make space for Vista. I believe there are a few files kicking around on the XP partition that are something to do with Vista so I won't be touching those and I've changed the boot.ini file so XP is the default choice of OS to boot after 5 seconds.
I'll let you know as time goes on if I run into any significant problems, I doubt I will since I seem to be blessed when it comes to OS installs. My boss just installed Vista on his machine which is identical to mine and ran into some issues with drivers not installing that didn't happen to me plus his machine scores lower in the Experience Index. I had no problems at all with my install, odd!
I would have to second the idea of using a seperate hard drive. First low RPM IDE drives are relatively cheap. Second hard drives are usually a good long term investment in terms of computer upgrading. I have personally made use of some hard drives for 10 years through 5 different systems. After it has served the initial purpose of being a Vista test bed then it can become a long term backup storage device.
From what your saying it will likely take at least a year for 3rd party components to make the transition and catch up from Delphi 7 to 2007. This means that either you are going to have to do a ton of work fixing and debugging other peoples code to make the transition or you are going to have wait for them to make the updates. In the mean time you need to have a stable development environment and be able to compile CMud. To my thinking this means it is best to have the least disruption of your current Dev system, the smallest disruption being no change.
Definitely moving ahead and having a Vista system and the latest in development tools will be advantageous. However, jumping the gun is not the right move. That lag of 6 months to a year while Vista becomes more stable, 3rd party components develop, and bugs in Delphi 2007's support of Vista are worked out will make a very large difference in both the quality of your end product and how much frustration you have in trying to produce it.
Which brings up another interesting point; until just a little while ago Vista was in beta. Meaning that any major support of it from Delphi 2007 was built during the Vista beta. With the likelyhood that there will be many changes to Vista within the first 6 months, I wouldn't trust Delphi 2007 to have it all totally right just yet.
Overall that makes me say wait 1 year until the next conference, when they are likely to offer the same types of discounts. Prior to that point there is far too large of a chance of instabilties in Vista, Delphi 2007, and all the 3rd party components causing problems that it isn't worht upgrading at this time. I do not advocate waiting 7 years. One is enough. If you really feel the need to push the edge and spend your time on it then please make it as isolated from your current Dev system as possible.
_________________ The only good questions are the ones we have never answered before.
Search the Forums
Joined: 25 Sep 2000 Posts: 23377 Location: Colorado, USA
Posted: Sat Mar 17, 2007 1:00 am
Actually, it's better than you think Vijilante. CodeGear did something very smart with Delphi 2007 (and very tricky from a technical perspective). They made Delphi 2007 binary package compatibile with the previous Delphi 2006 version. So any 3rd party that has Delphi 2006 components will already work with Delphi 2007. Also, with source code (which I have), it's a relatively straight-forward recompile of the components to get them into Delphi 2007. So as I mentioned, I think it will only take a couple of weeks...definitely not a year.
This isn't a "Vista test system". This is a "Vista development system". The goal will be to get the software that I need installed and working within a couple of weeks so that I can use Delphi 2007 and Vista as the primary CMUD development.
So, I don't want to use cheap IDE drives. I want the same RAID SATA mirroring that I had with WinXP. I need the data reliability (along with my regular backups) to avoid any disk problem issues like I've had in the past.
The main reason I'm making this move now is totally based upon the conversations I had with other Delphi developers at the virtual CodeRage conference. Both developers who have used Delphi 2007 already and Vista users. From the research that I did, it looks like the transition from Delphi 7/XP to Delphi 2007/Vista is pretty painless. I'll definitely let you know how it goes, but it does not sound like the huge effort and horror story that I originally thought it would.
Delphi 2007 is not a rewrite. It's pretty much the improved Delphi 7 compiler that they put into the new Delphi Studio 2006 IDE. Remember that the Borland focus was on .NET stuff for the past few years. So the Win32 stuff has been nice and stable. And the IDE was rewritten back in 2005, so they've had several years to improve the IDE and make it stable. Their stated purpose with the Delphi 2007 release was to fix all of the issues that were stopping Delphi 7 customers like me from upgrading in the past. The feedback from the D2007 Beta testers in the conference were more positive than I have heard in a long time.
Yeah, we'll find out for sure next week. Delphi 2007 shipped today, so I'll have it Tuesday. I'm installing Vista on it's own partition tonight.
But don't worry, I'm not going to let any of this screw up my current development system or take time away from working on new CMUD versions.
It was funny to read your post though...because that's exactly what I was thinking a couple of days ago. It was only after talking to some respected Delphi developers that I started to change my mind and become more positive about making these changes sooner rather than later.