Fog Creek Software
Discussion Board




"Designed for Windows XP" Logo - Is it worth it?

As an ISV, we have been invited by Microsoft to take part in its Logo program.

All well and good, but the requirements, such as placing some Registry values into Local Machine instead of Current User (for example) goes against the way we developed our product.

Looking at the CityDesk website, I am interested in the absence of any 'certification' logos, despite Joel's high standard of coding quality and product design.

I would like to know if Joel, and others here, are not interested in the logo program, because it is too much effort for too little gain?

Or are there other reasons?

Thanks in advance!

Andrew

Andrew Whitten
Tuesday, February 17, 2004

"...such as placing some Registry values into Local Machine instead of Current User (for example)..."

Have you got that the right way round?

John Topley (www.johntopley.com)
Tuesday, February 17, 2004

There are, presumably, some political reasons that it's helpful; namely, it may place you in good stead with the local or national Microsoft offices, which may give you new opportunities.

As for whether end users care? No, they really don't. Will some corporations only buy logo'd software? I suppose that's possible, but I've never heard of it.

Brad Wilson (dotnetguy.techieswithcats.com)
Tuesday, February 17, 2004

"Have you got that the right way round?"

The logo would mean something to me if it forced vendors to ensure their software didn't need to run as admin. That's one of my biggest gripes about Windows software in general.

Brad Wilson (dotnetguy.techieswithcats.com)
Tuesday, February 17, 2004

Agreed.

Didn't someone do a study once that concluded that Microsoft Office didn't actually qualify for the logo?

John Topley (www.johntopley.com)
Tuesday, February 17, 2004

When I ran our App through the Microsoft Verifier, I recieved this message, which describes my issue far better than I did:

"Designed for Windows Logo Requirement 2.8. The application wrote to the HKEY_CURRENT_USER area of the Registry. Writing to this area of the Registry at installation time prevents the information from being available to all users. Configure the application to write to the HKEY_LOCAL_MACHINE area of the Registry or an All Users data directory."

By the way, I agree that many Microsoft apps would not pass such a test.

Our customers are, however, typically government, and such a logo might be a deciding factor for them far more than a home user.

Andrew Whitten
Tuesday, February 17, 2004

Two questions here (actually)...

1) Do you need the logo?
2 Should the program be compliant?

It's the second question that is most important.

(No doubt, there are a "few" MS products that are not compliant but that fact is irrelevent and would be MS's problem to fix.)

njkayaker
Tuesday, February 17, 2004

I heard that one of the big office-supply stores was going to require it for printers and other (hardware) peripherals, because hardware companies are notoriously bad at software and uncertified hardware usually implies all kinds of installation problems and a higher than usual return rate.

I also heard that if your product is a cheap consumer application that you want to sell bundled with hardware, e.g. like an encyclopedia that you want to ship with IBM-PCs, certain hardware vendors, like IBM, will require the logo.

A third reason is that it's one way to get into various Microsoft ISV programs which can reduce the cost of MSDN subscriptions.

But for 99% of the software ISVs out there, it's hard to find a return for this particular investment.

Joel Spolsky
Fog Creek Software
Tuesday, February 17, 2004

Yep if you have MSDN licenses anyway then it's a great idea. Costs $1500 (IIRC) per year for either 5 or 10 (depending on level) developer licenses plus upgrade licenses for XP & Office along with server licenses. (Plus unlimited demo / training licenses)

Writing to HKLM rather HKCU at install time isn't a bad idea. We do BOTH and we passed testing. i.e. HKLM is the default value, HKCU is the current users value. We write a huge amount of stuff to HKCU at runtime, i.e. window positions etc but that doesn't count.

The worst part about this testing is removing yourself afterwards we ended up writing a little stub exe to remove all traces of ourselves from the registry (i.e. delete HKCU/Software/CompanyName/Product/* plus HKCU/Software/CompanyName if blank).

Basically it depends on how many MSDN subs you currently have. My guess is that it's an ideal task to get an intern to do most of the grunt work on (i.e. tweaking install scripts etc) and should take a couple of months.

Personally given that I used to have 2 full MSDN subs plus some upgraded copies of Visual Studio I can justify the expense just from the savings I made in that area (not upgrading from VS6->VS.net since I now have 10 full MSDN copies), plus I have a 100 user network license and office licenses.

As a guide on our first pass (for Verified for Win2K certification) we spent $1,200 plus $800 for a retest afterwards. We did the fuller "designed for XP gold" test a couple of years ago and this we passed first time at $3000. I have a suspicion we were the first ISV to go down this route as the MS enrollment web pages didn't understand about the test and we had to be pushed through by hand.

The older "Certified for Win2K" test was much harder and had things like self-repair, windows installer and forced use of HTML help and a few other things that were technically too much like hard work for us, so we didn't bother.

From the sales point of view it may help with some sales, we find it helpful (our app costs $1,500 per user) but it's not a deal clincher. It does help give an aura of being "solid" software but we just put the logo on some of the marketing material and don't push it.

Oh and the gold partner plaque this year looks like some money was spent on it.

Peter Ibbotson
Tuesday, February 17, 2004

I've had to deal with a lot of software that not only didn't have the logo, but it was quite clear that the developers didn't understand windows XP.  I too am tired of apps that require Admin rights to use some or all of their features.  Since part of my living comes from doing computer security evaluations, it's fair to assume that I'm not going to give my kid's account admin rights just so that he can run his favorite game.  Clients would start asking for their money back if they found out.

My feeling is that it is worth it, not from any value that the logo gets you, but because it means your app is less likely to bite somebody in the butt because of unforseen security issues.

Clay Dowling
Tuesday, February 17, 2004

If Microsoft were sensible, they would design the requirements for their logo program to make it easy for compliant software to be ported to future versions of their OSes.

But that's conditional upon Microsoft being sensible...

C Rose
Tuesday, February 17, 2004

"If Microsoft were sensible, they would design the requirements for their logo program to make it easy for compliant software to be ported to future versions of their OSes."

Care to elaborate?

John Topley (www.johntopley.com)
Wednesday, February 18, 2004

As far as XP is concerned they did do that we took our "verified for Win2K" software and ran through the tests for "designed for XP gold" without any real changes (I know there was one minor change we did but I can't remember what)
Other than that (IIRC) they force you as part of the test to change the reported windows version number. Software is allowed to barf on higher version numbers BUT only if it's something like backup or disk defrag software.
Generally all of the stuff in these tests are GOOD THINGS to be doing anyway.

Peter Ibbotson
Wednesday, February 18, 2004

Many thanks for all your comments!

My feeling now that it is worth it for the MSDN benefits, and because it is a guide to good Windows programming practices; even though it may be the start of the journey rather than the automatically gaining of the 'trophy' of a Logo!

But, as many of you have said, it is more 'icing on the cake' than an immediate benefit. I hope to steer our product towards this goal, and hopefully it should pass with much less effort by version 2.

Andrew

Andrew Whitten
Wednesday, February 18, 2004

*  Recent Topics

*  Fog Creek Home