Small Developer, Big New OS
Posted June 26th, 2003 @ 10:24pm by Erik J. Barzeski
I make no bones about the fact that Freshly Squeezed Software is a small company. We're three people, really. But (roughly) so are other great Mac software development companies, like Panic and even Omni Group to a certain extent. Other companies are only one person.
Given that we're small companies, and even though we have some pretty nifty software, we often don't have the resources (i.e. cash) to have fourteen machines laying around running various versions of the OS on different configurations. In other words, when Mac OS X 10.3 comes out, I will install it on my machine. Testing FSS software on 10.2? That ends at about the same time (or when the last of us upgrades, anyway).
We cost ourselves sales doing this. Heck, many people still ask us if our software works on Mac OS 9. No, and with the releases after Mac OS X 10.3's release, our software may not work on 10.2.x. We'll keep our last 10.2-certified release online for those customers yet to upgrade, but we simply don't have the time to support "older" OSes. We're moving forward, for the benefit of our users (better tools, cooler widgets and APIs, allow us to create better software), and we'd like our customers to join us.
It's a little bit of a sticky situation. MailDrop still supports Mac OS X 10.1, for example, because we didn't change much in the move from 10.1 to 10.2. But PulpFiction? MailDrop 2? They'll be 10.3 or later, most likely.
We can't afford not to adopt some of the awesome new things the Cocoa APIs in 10.3 will provide for us. Can we afford to lose a few customers who haven't upgraded yet to Panther? I guess. But at the worst case, those folks are still future potential customers.
As a small development company, we can only do what we can do.
Donate Life
Posted 27 Jun 2003 at 2:37am #
Make stuff that makes you happy. If using all the fancy new bells and whistles makes you happy, go for it. That's how I code.
Posted 27 Jun 2003 at 5:18am #
Any reason not to go pickup an old G3 beige and slap 10.1/10.2 on it? Other than the $500 it might set you back? Dunno your sales figures, but you might recoup that if you can make the current stuff run on older platforms.
i say as I write in 10.2.6...
Posted 27 Jun 2003 at 8:00am #
I have three machines, but one typically is ignored (my portable), and one is my "forward" machine (test seeds of the next OS), and one is my "working" system - the current OS, the one on which I write email, etc.
I could spend $500, but frankly, then I'd need a display and the space to keep the thing. If I were making a living off of my software (and the Panic guys are, and certainly OmniGroup, and some others, and I certainly aspire to get there some day), then I'd buy a machine for these purposes. Until then, I remain a "small" developer with an already too-large electric bill and an already too-small apartment (computer books, computers, and equipment and gadgets take up a lot of room!).
Posted 27 Jun 2003 at 10:42am #
I think it wouldn't be a problem, getting another disk and running 10.2 and 10.3 on the same machine.
But developing for two operating systems is quite difficult. You couldn't use new fancy stuff of Jaguar or you have to develop two different versions.
Anyhow, the majority of MacOS X Users will switch to Jaguar very soon, so the loss of customers wouldn't be so big.
And for those witd the 'old' OS, they could use the 'old' version of the software.
Posted 27 Jun 2003 at 11:30am #
Hey, we're right there with ya. There's two of us and we have the same problems and we came to the same conclusions that you did. The biggest problem for us isn't support of the OS it's support for hardware types.
Since our software, Fluid in particular, is usually heavy on the graphics end it can be extremely hard on the laptops out there. Do we have three or four different PowerBooks to test on? Nope. We can't even afford to buy one of them.
Of course, most people think that as long as a program is popular that the developers must be making money at it. That's quite funny. I think we've made just over $400 dollars from Fluid which is one of the most popular screensavers for OS X. Granted, we released it without the need or preconception of money earned (it was really just to get our name out there, and it has succeeded in that). But, it is interesting to see the realities of shareware payment at work.
I'd be interested to know what other shareware authors see as far as download to payment ratios...just the geek in me wondering.
I think if the general public realized the incredible amount of work and time involved in writing quality software they would probably understand the position of our companies with respect to limited platform choice (which includes all those idiots that write nasty letters because we won't port Fluid to Windows).
Posted 27 Jun 2003 at 12:30pm #
Addition for your list of small developers: Brent Simmons of NetNewsWire... yet another husband/wife team doing a good job of independent software development.
Here's a question: Do you have a feel for how much money you're leaving on the table because you are supporting only the latest and greatest OS? Would it be worth some small amount of money to have an independent team test? Or maybe rent a lab full of different hardware/software configurations for a week?
Posted 27 Jun 2003 at 2:15pm #
I say yes and no on the independent testing. It might make it easier to make sure nothing is broken if the older system is being targeted. It won't be helpful if we're moving to a new system to take advantage of the new frameworks. When 10.3 comes out, it's going to have a ton of new features in the frameworks that are going to be really hard to forgo just to keep older versions of the OS alive.
Posted 27 Jun 2003 at 2:28pm #
I think my new philosophy is develop under the latest and greatest OS and test under an older OS if possible. Since I've heard rumblings that Panther won't run on a beige G3, that machine just became a lagacy test machine.
If something doesn't work and it's requires a minor fix without sacrificing functionality, then I'll try to fix it. If the changes would result in diminished functionality, too bad.
When the G3 kicks the bucket, then no more legacy testing... at least until something else I own will no longer run the latest and greatest.
Course, since I never actually finished any major projects under my old philosophy, you can take this with a grain of salt if you'd like 🙂
Posted 01 Jul 2003 at 2:29am #
Just got a ADC-mail. XCode is available for download now. [link] And, what is important for this comment it is possible to build for X.1 and X.2 also: SDK Support - You can now develop software for Mac OS X 10.1, Mac OS X 10.2 or Panther Pre-release from one machine
Posted 01 Jul 2003 at 5:07am #
XCode
At the WWDC Apple anounced XCode. XCode is a new integrated development environment for MacOS X. It has some great new features like 'zero link' or 'fix and continue', which allows faster debugging cycles. That are real helpful things, because...
Posted 01 Jul 2003 at 6:58am #
I don't think it's available for download. It says it's on the July mailing, which I imagine I'll get in the next day or two.
And even though it will build software that runs on 10.1 or 10.2 - of course it will - the point is that we have little or no way to test the software because we're running 10.3. There are other factors as well, such as really needing a feature the 10.3 frameworks provide, and so forth.
Posted 01 Jul 2003 at 7:35am #
Yes, you're right - no download.
But I don't see the testing problem. Setting up a second disk or another partition on a disk shouldn't be a problem.
Developing for different SDK is in deed difficult, but doable if you use OOP, a good system design and a versioning system with branches.
If it is worth this effort? Don't know; depends...
Posted 01 Jul 2003 at 2:07pm #
I was debating this all week at WWDC, and I know others feel the same based on the Q&A segments after various talks.
For the moment, my app is small enough that it's practical to keep the Jaguar codebase and fork off a Panther version with more features (and likely smaller codebase). I can keep the same model objects, of course, which is the only thing that makes this approach plausible.
The only thing I've decided on is that I will be taking advantage of Panther features in some form. That may mean maintaing two separate projects. It may mean releasing a single Jaguar version and focusing major additional development the Panther version.
A lot of this will be decided by how well Apple is able to communicate the value of Panther to end users. I'm also a little concerned if it will be clear to people that Panther is the successor to Jaguar. I find myself mixing the names up soemtimes. Pentium IV is devoid of creativity, but it is at least clear that it's the newest iteration. It would be nice if there was something in between.
- Scott
Posted 01 Jul 2003 at 6:01pm #
Apple's naming scheme is a bit confusing, as I sometimes hear people say things like "I have Mac OS X 10.2, do I need Jaguar?" There's a disconnect between the two - the name and the version - that confuses some.