Mar 93 Editorial
|Column Tag:||The Editor's Page
The January Pilgrimage to Macworld
By Neil Ticktin, Editor-in-Chief, Don Bresee, Technical Editor
Since the beginning of time (which of course was January 21, 1984), the Macintosh community has made two trips per year to the grand-daddy of all Macintosh trade shows - Macworld Expo. The first of this years Expos was held January 6-9, 1993 at the newly expanded Moscone Center in San Francisco.
Aside from having the Expo on about the worst date possible, the show went off better than usual. It seems that the expo coordinators are not aware that many, if not most, Macintosh related businesses (including Apple) are closed the week between Christmas and New Years. This increases the difficulty for most people to make the show.
The good news is that San Francisco has completed the long awaited expansion of the Moscone Center. For the first time in many years, the entire Expo was held in one place, eliminating the annoying bus ride between locations. Congratulations to all involved, this made the show much nicer to attend.
The Hot Item - Components
Overall, the show was solid, unveiling many improvements to many things, as you might expect. One product was the buzz of the developer world - Component Software Corporations new product, Component Workshop.
Some of you might remember back from your computer industry history, On Technology was formed several years ago by Mitch Kapor. Originally, On was going to work on a fully object oriented, portable operating system (similar to Taligent). After some time, On became more realistic in their goals. A good move because they would have had to go head to head with Taligent. On decided not to work on this product hence the new company - Component Software Corporation (CSC).
What is Component Workshop? According to CSC, Component Workshop (CW) is the first dynamic C++ development environment, targeted for new application development. Most object-oriented systems have derived properties from Lisp. The difference between CW and other implementations of C++ is that, like Lisp, CW doesnt separate the development environment from your application. Instead CW uses an Extruder to analyze your application and generate C code.
Developing with CW is different from traditional approaches. Instead of developing the entire application at one time, you develop it one part at a time. Using traditional classes and subclasses, your work becomes reusable by other projects, at your convenience.
CW also differs from C++ in that it deals with all memory management. Instead of traditional, compile/link cycles, CW uses an incremental linker and compiler to constantly keep your project up to date. This is all done through its dynamic environment.
I could go on about Component Workshop, but that would be better done in a formal product review. The bottom line is this: there are a lot of people moving towards CW including most notably, Bowers Development. Not only will AppMaker (from Bowers) be supporting CW, but the next release will be written in it.
Component Workshop running on the Macintosh platform is shipping and they are slating a Windows version for the third quarter of 1993. CW is going head to head with Bedrock. Bedrock is backed by Apple and Symantec. On the other hand, CW is out now.
Component can be reached at 617/862-9700 or Fax: 617/862-7749.
Covertest from Jasik
Steve Jasik (of MacNosy/The Debugger fame) has released a new, simple tool to determine code coverage in a program. CoverTest helps the tester (or programmer) determine which parts of a program have been executed. It extracts symbol information from MPW and shows the information in bar graph form.
The product will be bundled with The Debugger and has been shipped to all owners of the current version. For more information, contact Jasik Designs, 415/322-1386 or you can order The Debugger (with CoverTest) through the MacTech Mail Order Store.
Apple is now showing MPW 3.3. The beta for this version will ship on E.T.O. #10 in February, 1993. The final version is estimated to ship in April/May 93.
There are several new features in MPW 3.3, the most important of which is the new incremental linker. This allows MPW to link only those files that have changed. This should greatly reduce your link time after a simple change. MPW will reportedly link in less than a minute even for complex applications. The linker does accept the same command line options as the current linker and does work with current build scripts/make files.
There are a few limitations to the new linker. For example, it will only link applications and tools. It does require more memory (up to 3x the size of all object files) and more disk space. You do need to use the normal linker for your final build and the incremental linker does not produce .SYM files
MPW 3.3 also comes with new documentation. There will be three new books: Introduction to MPW, Building and Managing Programs with MPW, and MPW 3.3 Command Reference. Reportedly, these documents are completely rewritten, reorganized and incorporate existing release notes.
Version 3.3 also comes with other improvements including better System 7 support (i.e., Alias resolution in all MPW tools and Apple Event support). There is keyboard navigation in dialog boxes for dialogs such as Projector and Find/Replace. And, Command-P now is for Print too. Finally, there are many bugs fixed overall, but specifically in Projector and Object Pascal.
Apple will be shipping ToolServer 1.0 and SourceServer 1.0 in this version. ToolServer is an AppleEvent aware MPW Shell which allows MPW users to have tools and scripts run on a remote machine. It does allow 3rd-party environments to access MPW Tools and Scripts. Reportedly, THINK C and Pascal will do this in an upcoming version. Support is also expected for ACI USs Object Master and Calliope.
The Publisher's Column
By David Williams, Publisher
Over the holidays, I went to Europe. To me, one of the most fascinating things about Europe is how the people approach computers. I went in and out of several stores in Italy, France, and Switzerland, and talked to numbers of users. What I heard got me thinking about the direction that Apple has decided to take, and, as usual, I hope to raise some points here that will provoke a response.
Europeans dont like lots of options on the list. They like a fairly small number of models to choose from, and they dont respond well to models that change often. A new car in europe often sells best in its third year, and rarely sells as well in its first. New technology is welcomed, but preferably as improvements added to existing models, so its impact can be clearly seen. Anyone familiar with with the Mercedes-Benz series of cars will recognize these traits in their line.
Motorola recently announced the 68060 chip, which contains 2.2 million transistors and will initialy run at 50MHz. Intels Pentium chip contains 3.2 million and will initially run at 66MHz. Meanwhile Apple and IBM are fast at work on PowerPC. The 601 might actually be available from IBM in late 1993. The opportunity for new models and features is clearly the highest its ever been.
In response to this situation, Apple has announced a new strategy of rapid turnover of models, so that the most up-to-date technology is available at todays price - especially in the PowerBook line. The theory is that the buyer should be able to buy the most computing power for the money, right at that moment.
From the perspective of a purchase decision-maker for a medium sized company (my largest client) I find this new strategy very confusing. To explain, I have always viewed Apple as alone, apart, but stable. The Mercedes Benz of computing. Macs were very expensive, but didnt go obsolete very often. Instead, upgrade paths were available, and existing Macs were usually able to take at least partial advantage of new developments. Most important, with a Mac, one very rarely heard employees complain that their work went slowly because their computer didnt have a large enough cache.
One of the advantages of the Mac was that there werent a plethora of models. With only a few to choose from, you didnt have to worry much that a supervisor in department A had a faster machine than one in department B, prompting B to ask for a math coprocessor and a caching controller.
PCs on the other hand were cheaper, but inherently evanescent. The hottest thing in town one month was old junk in a year. This situation was entirely acceptable however, because it was the result of competition between the many, many clones. As soon as another clone one-upped your new machine, you had to expect your vendor to respond.
This was in no way similar to the auto-makers. The newer machines were actually better, not just longer, lower, and wider. No one ever thought that Compaq brought out a new model just for a new model year. On the other hand, employees did want to buy each and every new enhancement that came out.
Corporate buyers were able to say to their boards, Yes, the Mac is expensive but it wont be obsolete next month..... Now, Apple plans to introduce new models, especially of notebooks, with great speed. Apple will not wait for new chips, but will instead provide an ongoing series of constant small improvements, interspersed, of course, with large ones. Needless to say, this is somewhat of a complete reversal from the days when one could count on a Mac to remain a current model for a fairly long time.
Unlike the clones, Apple cannot blame constantly changing models on direct competition. Instead, each new Mac must stand on its own merits, with corporate buyers such as myself asking why the change was necessary. One of the most difficult things to explain is why the newly promoted manager of department C has a new Mac that outperforms the Presidents one year old one.
With the 68060 available, in both low-power and regular versions, on top of the 68040, and the PowerPC chips as well, Apple will have the ability to bring forth a huge array of new machines. With ongoing changes in storage technology, display capabilities, and accesories, this array can change almost monthly.
If that is to be the case, I believe that Apple must maintain clear upgrade paths at affordable prices. Currently, the price of upgrading is far too high. It shouldnt cost nearly 50% of the original purchase price to upgrade a Quadra 900 to a 950 when the 900 was only sold for a few months. Further, knowing that the 950 will be made obsolete (or will at least be outperformed) by the 800 which will cost less, the 950 price has to come down from the beginning, or buyers will wait for the 800. In other words, more models faster means lower prices sooner. Finally, Apple will have to learn to turn out full production of a new model without the constant supply shortages that plague us today.
Clearly, consumers will benefit from more available technology at lower prices. Software developers should benefit from the greatly expanded markets, and after-market suppliers might benefit from supplying upgrade kits cheaper than Apple. It seems that all of the risk and most of the burden falls on Apple itself.
The question remains whether Apple will prosper with the change in its line and in its image. The US and European automotive industry has taught the lesson well: given the choice of being Daimler or GM, its better to show a profit on the same old models year after year, than show a loss on a huge number of new and improved ones.