I can't confirm nor refute Jesse's numbers or logic. However, I think there additional considerations.
I would think that one of MM's goals is to have CF be the most widely-accepted Web Application Server regardless of hardware -- comparative versions of CF Cost about the same regardless of platform. I am not claiming that OS X has enough potential to be a major player in the Web Application Server market -- what I am saying is that it should be a player, a profitable player! Consider this: There are lots of web developers out there (myself included) that use Macs. Many develop applications that are hosted by outside services. Those who develop in CF develop apps that will be run on CF servers on some platform. Point 1 -- These Mac developers contribute to CF sales Many Mac developers do *not* write Web programs for CF because they cannot run CF on a local machine -- they are more productive in Java, PHP, Perl, whatever. Point 2 -- These Mac developers are contributing to the sales of MM's competitors The port of the Developer system to Mac OS X alleviates this problem to some extent -- Likely, independent contractors like myself will jump at the chance to run CF locally. This should increase their productivity and they should be able to deploy more apps to supported CF platforms. Point 3 - Availability of an unsupported developer system on OSX will likely make a contribution to CF sales and detract from competitive sales The fact that the port is unsupported means that it takes a lot of cajoling, wheedling (not bribes), and time (especially time) to get problems fixed, or even addressed. Likely, there will not be mass acceptance of CF as a development platform, by Mac users, until there is a supported developer platform -- CF running locally, not just DWMX* * It's ironic that DWMX on the Mac contributes to the sales of CF competitors -- Mac users can develop in competitive languages that *do* run on the Mac, while CF does not. Point 4 - Lack of a supported CF Mac developer platform will likely detract from CF sales and contribute to competitor sales. Those who have used CFMX on Mac OS X (I know most of them) think that is a sweet system -- superior to the other available options. I can't make this statement, because I have never developed on a Win, Solaris or Linux box. I can give this opinion: CFMX ON Mac OSX is the *Best* Web Application development system, running on the *Best* Operating system running on the *Best* Personal Computer. Many will agree with me. Some simple test results, a few discussed in recent threads, show that that Mac platform performs quite well, and is price-competitive, if not superior. Point 5 -- Likely MM is missing an opportunity to proselytize Mac developers to CF, to their detriment and to the advantage of their competitors. If MM were to offer a supported Developer version of CF on Mac OS X, it would likely be the easiest, and fastest to install. Mac OS X already includes (has installed), Java and Apache Web Server. These are not necessarily installed on the CF supported platforms. A good part of every CF install (and many of the problems during beta) involve installation of Java, a Web Server, and integrating these things with the CFMX/JRun systems. Mac OS X has a proven, well-defined, automatic process that maintains and applies software updates -- this include everything from updates to the underlying OS to 3rd-party application programs such as Microsoft IE. I suspect a CFMX product on Mac OS X would cost MM less to support and maintain, than on other platforms -- there is just a lot less code and a lot less integration -- a complete, stable, predictable platform. Point 6 -- Costs of supporting CF on the Mac OS X platform would likely be less than other platforms. This means different numbers would be plugged into the ROI calculation. If MM offers a Developer version of CFMX on Mac OS X, they will likely find an eager audience of potential *new users* of CF. Apple will likely promote the availability of CFMX Developer on Mac OS X. MM could defray much of the marketing costs -- others will gladly do it for you. This might be a simple, inexpensive way of "sticking your toes in the water" to measure acceptance and potential without incurring a lot of costs. Point 7 -- Costs of marketing CF on the Mac OS X platform would likely be less than other platforms. This means different numbers would be plugged into the ROI calculation. Given, a supported CFMX Developer version for Mac OS X, there would likely be much pressure to release a supported CFMX production version(s) --- and the problem with this is??? These sales would be *New Business" full price sales, not upgrades from prior sales. Again, emergence of this "Full" product would realize the same (if not more) benefits of the Developer product on Mac OS X -- lower maintenance and support costs; lower marketing costs; a potential customer set that is willing and able to buy; likely heavy promotion by Apple; the *buzz* of the industry in all the trade pubs; a plethora of new books, articles, training, certification and other services. Point 8 -- Likely MM should evaluate CFMX on Mac OS X as they would any new product in a new market. There is a potential here, to create, and dominate a very profitable market. Sometimes you have to not go by the procedures, take the risks, do the best justification you can -- and then take action If Jeremy Allaire, hadn't done so, we would all be doing different things. Well, Martin Luther stopped at 10, so I better stop at 8 Dick Many of them do not use CF because they cannot develop on a local machine On Wednesday, July 31, 2002, at 10:27 AM, Jesse Noller wrote: > You just said the magic words! > > "niche market" > > Here's the problem. Most people assume (as did I a long time ago) that > porting to new platforms and operating systems was a simple matter of > recompiling the binaries, and BAM! "You've got Software". > > However, as this was the way I thought a few years back, I have since > had my eyes opened. > > Generally, when we evaluate a new operating system, or distribution, we > have to say: > > 1: How many people are going to buy said product for that platform. > A: How many users have requested it? > B: How many large corporations and businesses are running it? > C: How many existing platform users are likely to purchase a port? > > Let's say it breaks down to 100,000 users. Let's say we know of 3 > companies willing to buy the full product of a port. Let's say of those > 100,000 users, 30 users have requested it, and the market shows a > "strong interest". We can then assume (within reason) a sales market of > a few HUNDRED. Not thousands of users. Not with an initial port, and > zero market/platform penetration. > > Therefore, let's assume 10 companies, 500 users. This is a > conservative estimate... > > At say, 700$ break-in price for the base product, we can "assume": > > 510x700= 357,000$ > > Now, that can be our initial "money" we would garner from a port. That > would be the ROI (return on investment) for a product. > > 2: Let's now figure out development costs. Better yet! Let's base this > off of porting to OS/X. > > A: Servers/Machines minimum of 6 for development, minimum 6 for > QA, minimum 3 for technical support. > > At say, 5000$ a piece for each machine, that 75,000$ up front. > > B: Integration into the existing environment (IT misc), assume a > cost of 200$/hour (salary say) minimum 10 hours. 2000$ > > C: Training for existing developers, technical support and QA, > let's assume 4000$ a class, or misc costs, (not including man hours). > Let's just make it an even 20,000 cost. (uber-conservative). > > D: Development man hours: let's assume that each engineer is > making 30$/hr. Let's assume that minimum dev time for this is 1 month. > Working only weekdays. Let's assume 4 engineers. (eight hour day) > 19,200$ > > > E: QA Man hours... Being a member of QA... well, let's say this > estimates are conservative. Let's just make them the same as the > engineers (HAHAHA), 19,200$ > > [Please note, my math my be screwed up, I'm hopping around between > projects] > > F: Documentation! Let's just assume another 19,200$ cost. > > G: 3rd party licensing. Assume 50,000$ here, just to be safe. > > Total for basic Dev/QA: 204,600. > > Take the ROI, subtract the Dev cost: 152,400$ left over. > > Now, subtract Market training and development. Market research and > penetration research. Sales research and training. Other application > integration testing and design (Flash gateway, dreamweaver, etc). Take > into account Support training. > > Pretty soon, you're making nothing. > > > I had to do a lot of this type of research when pushing for added Linux > distro support. As well as BSD and other Unixes support. > > Sure, on the outside, the porting process seems pretty easy. Trivial in > fact. In reality, this is a software development firm, and we cannot > afford to do anything half-way. > > We can't give everything away and we can't be loose with our release > guidelines. For every single platform support we need a guarantee that > our ROI will, if not defraying the immediate costs, in the long run > outweigh the costs of development. > > Someone mentioned HP. Yeah. CF runs on HP. You have no idea how much of > a nightmare it is to try to test new-gen software on hardware that is > nearly 6 gens behind. > > Remember, we also have to pay upkeep, etc. The problem with niche > markets, is that while yes, it would give us some revenue, and it would > provide us with more market penetration, the problem is, will it assist > and defray the cost of us developing, supporting, etc the platform. > > Ok, back to more work. Woo. > > Jesse Noller > [EMAIL PROTECTED] > Macromedia Server Development > Unix/Linux "special guy" > >> -----Original Message----- >> From: Ian Lurie [mailto:[EMAIL PROTECTED]] >> Sent: Wednesday, July 31, 2002 12:20 PM >> To: CF-Talk >> Subject: RE: X-server? >> >> I don't know about offensive but it seems like Macromedia may be >> missing >> an >> opportunity here. The X Server is a screaming deal - inexpensive, fast >> and >> easy to work with. If CF MX worked on it it'd be a great niche market >> that >> you-know-who would have a hard time penetrating with .Net. >> >> -----Original Message----- >> From: Cary Gordon [mailto:[EMAIL PROTECTED]] >> Sent: Wednesday, July 31, 2002 9:11 AM >> To: CF-Talk >> Subject: Re: X-server? >> >> >> Speaking for myself, I find this really offensive (the idea of paying a >> Macromedia employee, not you Dick). Between my company and our >> clients, >> we >> give Macromedia enough money to by a room full of Mac boxes. >> >> Cary >> >> At 07:38 PM 7/29/2002 -0700, you wrote: >>> Well, MM's Unix/Linux "special guy" says he will do an Apache thingie* >>> for for OS X if offered a simple bribe -- an OS Serve box and a 17" >>> flat >>> panel display (I think he should hold out for the 22" studio display) >>> >>> Anyone want to contribute -- I'll pledge $500 towards a bribe! >>> >>> * Note "thingie" is a highly technical C++ programming construct. >>> >>> Dick >> >> >> Cary Gordon >> The Cherry Hill Company >> >> >> >> > ______________________________________________________________________ Signup for the Fusion Authority news alert and keep up with the latest news in ColdFusion and related topics. http://www.fusionauthority.com/signup.cfm FAQ: http://www.thenetprofits.co.uk/coldfusion/faq Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/ Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists