A couple of polls have been posted lately asking what your Delphi IDE you use most for development.
Delphi 7 is well in the lead at based on the profile details at Delphi Praxis. It looks like this only includes about 17% of the membership. If you're a member, you might want to update your profile.
The front page poll at CodeGearGuru shows Delphi 2007 slightly in the lead over Delphi 6/7. If you haven't voted yet, maybe stop in and let your voice be heard.
Monday, September 17, 2007
Wednesday, September 12, 2007
Is software assurance worth while?
CodeGear released RAD Studio 2007 and has announced new pricing. This pricing and their faster release schedule has helped spark some discussion about software assurance (SA) and whether a subscription is worth the cost and risk.
SACalc is a simple application to compare the cost of purchasing Delphi upgrades against SA over time. The prices shown are for RAD Studio 2007. You can find prices for BDS 2006 and Delphi 2007 for Win32 on the shopping site. If you have pricing for older versions, including SA, I'd like to see that, too.
Buying Architect without SA, you pay $3299 up front and $2499 when an update is available, which means you will have spent a total of $5798 by the first upgrade.
Buying Architect with SA costs $4289 ($3299 + 990) up front and $990 annually, totaling $5279 after the first year. This is $519 less than if you were purchasing upgrades.
With 12 months between releases, SA costs less than purchasing upgrades in 8 months for the Architect SKU and 10 months for Enterprise. It takes longer to break even with Professional, at least partly because upgrade cost is proportionally lower compared to the other SKUs. In case someone doesn't like the monthly break down, I have also included the number of upgrades required before breaking even.
For existing customers, the results and savings are the same if you start with an upgrade instead of a new purchase.
So, is software assurance worth while? From a strictly financial point of view, if you intend to upgrade frequently, software assurance costs less. I think it's safe to call it a good deal.
If you find any problems with my assumptions or the code, please let me know.
[Update]
I followed a suggestion and changed "Months" to read "Months between releases" to make its intention more clear.
I defaulted to 12 months between releases to match the SA period, which is renewed annually. If you take the average time between all previous releases of Delphi, it comes out to just under 14 months. This isn't a bad number to use for your calculations, but I'm not sure past performance is necessarily indicative of future releases. RAD Studio 2007 (Highlander) came out 6 months after Delphi 2007, and if the planned releases in CodeGear's road map are even close, the next couple (Tiburon and Commodore) will be 6-9 months apart. This only takes Delphi in to account, not C++.
Not that I think anyone should buy SA based on one or two releases. I think it works best in the long term. Hoping to get a single additional upgrade out of it is just gambling.
SACalc is a simple application to compare the cost of purchasing Delphi upgrades against SA over time. The prices shown are for RAD Studio 2007. You can find prices for BDS 2006 and Delphi 2007 for Win32 on the shopping site. If you have pricing for older versions, including SA, I'd like to see that, too.
Buying Architect without SA, you pay $3299 up front and $2499 when an update is available, which means you will have spent a total of $5798 by the first upgrade.
Buying Architect with SA costs $4289 ($3299 + 990) up front and $990 annually, totaling $5279 after the first year. This is $519 less than if you were purchasing upgrades.
With 12 months between releases, SA costs less than purchasing upgrades in 8 months for the Architect SKU and 10 months for Enterprise. It takes longer to break even with Professional, at least partly because upgrade cost is proportionally lower compared to the other SKUs. In case someone doesn't like the monthly break down, I have also included the number of upgrades required before breaking even.
For existing customers, the results and savings are the same if you start with an upgrade instead of a new purchase.
So, is software assurance worth while? From a strictly financial point of view, if you intend to upgrade frequently, software assurance costs less. I think it's safe to call it a good deal.
If you find any problems with my assumptions or the code, please let me know.
[Update]
I followed a suggestion and changed "Months" to read "Months between releases" to make its intention more clear.
I defaulted to 12 months between releases to match the SA period, which is renewed annually. If you take the average time between all previous releases of Delphi, it comes out to just under 14 months. This isn't a bad number to use for your calculations, but I'm not sure past performance is necessarily indicative of future releases. RAD Studio 2007 (Highlander) came out 6 months after Delphi 2007, and if the planned releases in CodeGear's road map are even close, the next couple (Tiburon and Commodore) will be 6-9 months apart. This only takes Delphi in to account, not C++.
Not that I think anyone should buy SA based on one or two releases. I think it works best in the long term. Hoping to get a single additional upgrade out of it is just gambling.
Saturday, September 1, 2007
Highlander beta blogging: VCL.Net and Vista themes
You may have seen this image from Steve Trefethen where he demonstrates how the VCL in Delphi 2007 supports Vista Themes out of the box.
Highlander features an updated version of Delphi for .Net, which supports .Net 2.0 and also handles Vista themes properly.
Highlander features an updated version of Delphi for .Net, which supports .Net 2.0 and also handles Vista themes properly.
Highlander beta blogging: Help improvements
I have been given permission to blog about the Highlander beta.
There has been some discussion on the challenges of managing the new help system and some of the improvements we can expect to see. I especially liked Dee's comment about restoring code examples and including them in the automated testing suite.
Well, the proof of the pudding is in the eating. From using the help in Highlander, I can say:
- F1 is more reliable. I seldom see "No help found for context." messages any more.
- The help itself is a separate install to make future help-only updates easier.
- Code examples have returned. I saw this updated code sample for TClipboard.Assign:
CodeGear is delivering on the "Continuously improved and frequently updated documentation" item from their road map.
Of course, the help still isn't perfect. If you see any problems, feel free to use the feedback link found at the bottom of each help topic to let Codegear know.
There has been some discussion on the challenges of managing the new help system and some of the improvements we can expect to see. I especially liked Dee's comment about restoring code examples and including them in the automated testing suite.
Well, the proof of the pudding is in the eating. From using the help in Highlander, I can say:
- F1 is more reliable. I seldom see "No help found for context." messages any more.
- The help itself is a separate install to make future help-only updates easier.
- Code examples have returned. I saw this updated code sample for TClipboard.Assign:
CodeGear is delivering on the "Continuously improved and frequently updated documentation" item from their road map.
Of course, the help still isn't perfect. If you see any problems, feel free to use the feedback link found at the bottom of each help topic to let Codegear know.
Subscribe to:
Posts (Atom)