Calculating the Cost of Software Assurance
Bruce McGee is a nice, level-headed guy. He’s always interesting to read in delphi.non-technical, and I always appreciate his clear-thinking posts. Over the years, Bruce has given us tons of hard-hitting but constructive criticism.
Lately, he’s been, like me, a proponent of Software Assurance. Bruce is a subscriber (thanks, Bruce!) and it really works for him. He’s been discussing it with folks, and came to the conclusion that some folks weren’t totally clear on the benefits. So, being Bruce, he did something about it.
He’s posted for your use an SA Calculator that you can use to figure out if SA works for you. My guess is that it will — we are working hard to add real, obvious value to SA customers.
SA might not be for you, but please, at least make the decision with the numbers in front of you. 
I decided for my boss to buy 10 SA from years.
September 13th, 2007 at 11:27 amIt’s a protection for our Delphi investment and a cheap way to have last versions always on (or in the box ready to install as happens with delphi 8). It could appear as a bad investment but we save a lot of money and delphi ide is better than in the past.
Some work is needed but new commitment for developers is very good.
Delphi rocks!!!
SA is the only means of receiving support!
September 13th, 2007 at 1:39 pmThe numbers there assume that you must buy every new release, which in my experience few people actually do. Especially when the "upgrade" is of the quality of Delphi 8 or BDS2005. So based on our having 5 copies of D7 architect we would need to be paying less than $400/yr for SA to make it worth having, since BDS2007 is the first upgrade available from there.
We are still waiting for the two key features that would cause us to buy a new version - 64 bit and unicode. We actually have one copy of SA just so we can try the new versions, but we have not even been able to build our major project in anything between D7 and D2006.
September 13th, 2007 at 2:51 pmBruce has done a nice bit of work here, but his default value for the cryptic "Months" edit field is wrong.
In the NG discussion we had on this subject he pointed out that the average release gap for Delphi has historically been 14.5 months (history being a more reliable indicator than any future, non-committing promises).
Plug 14 months into the calculator and Pro SA doesn’t save money for 4 years!
September 13th, 2007 at 4:21 pm… and of course Moz is right. SA makes most sense if you take and use every upgrade.
In fact for Pro, SA ONLY makes sense if you do this. Plug 24 months into that months field (i.e. representing only every other update being of usefulness) and watch Pro ROI hit "None" across the board.
The problem being that you cannot know if an update you will get at some point in the future with SA will be of use unless you pay for it months in advance.
An Enterprise user that only ever takes (makes use of) alternate upgrades still comes out better off with SA in the long run (> 5 years though). An Architect user "wins" after just 4 years of using only alternate upgrades.
For a Pro user this is not the case.
Even just one release of no utility to a Pro SA user will make it very difficult for SA to ever "pay for itself".
It’s good to have the numbers in front of you, but a _Pro_ user still also needs a reliable crystal ball in order to make sure they are using the right numbers. Using what turn out to be the wrong numbers would cost you dearly.
Nice work Bruce, and qudos for Nick for bringing this tool to peoples attention.
September 13th, 2007 at 4:52 pmBanking on every second update being worth while doesn’t see likely to pay off to me. Historically Borland led from the front but still had significant gaps between must-have upgrades - D3 worked, D5 was usable but not essential, and I (like many others) went from D3 to D7 in one step. I would have gone to D6 but we waited a few months and got D7. Add in that CodeGear are now following the pack on features rather than leading, and it seems even less useful - if you want to be on the bleeding edge of the technology, you’ll be using VS. So from my experience, every 3 or 4 versions is fine.
I’m really hoping that BDS2008 is usable, and that the promised new features actually arrive for BDS2009, but after seeing 64 bit left as vapourware I’m not holding my breath. I suspect that by 2009 we will be down to one or two Delphi developers with the rest using VS. But we will have more than paid for those two copies of BDS2009 by keeping one edition of SA current for the intervening years. Then we’ll have to buy an extra license.
September 13th, 2007 at 8:42 pmThanks for the plug, Nick.
From Joylon’s suggestion, I changed the label from "months"
to read "Months between releases".
I updated the article with some thoughts on the time between releases.
For the record, I have owned every version of Delphi and used each of them in production, including the "dreaded" versions 4 and 8. The only one I found to be unusable was the unpatched Delphi 4 (the patched version works just fine).
September 14th, 2007 at 4:41 amThanks for the plug, Nick.
From Joylon’s suggestion, I changed the label from "months"
to read "Months between releases".
I updated the article with some thoughts on the time between releases. I fugure my speculation is as good as anyone else’s.
For the record, I have owned every version of Delphi and used each of them in production, including the "dreaded" versions 4 and 8. The only one I found to be unusable was the unpatched Delphi 4 (the patched version works just fine).
September 14th, 2007 at 4:42 am