I completely forgot to put this in even though Mark reminded me of it. March 17th marked the first annual SPUD, or Shareware Pay Up Day. On SPUD, you go through your software collection and send in all outstanding shareware payments to those dedicated programmers who provide us with excellent programs. In honor of TidBITS' upcoming one year anniversary, I encourage you to send in shareware payments, postcards, or whatever the author asks for. If you have a lot of shareware and can't afford to pay for it, at least send the authors postcards (19[cts] these days in the US) thanking them for their programs and telling them you'll pay when you can.
Glenn Fleishman of Yale University Printing Service writes about Multiple Master from Adobe, "It will not have serif-to-sans-serif masters! You're thinking that this will be like Donald Knuth's TeX thing, Metafont, where all aspects of a font are attached to dials. Multiple Masters will have a normal light, a normal black, an extreme condensed, and an extreme expanded master in each font. By twiddling dials, you can get, for Futura say, Futura Light Condensed, Futura Regular Bold, Futura Expanded Light, etc. But in no imaginable universe take Univers and twiddle a dial and get Univers Serif Roman. Hermann Zapf designed Optima to be a serif face without serifs (i.e., thick and thin strokes, instead of more uniform strokes); how would you turn a dial, and zip-zip-zip, get Optima Serif? I'm not really outraged; I'd very much like a program or utility that did that. But see Douglas Hofstadter's discussion of Metafont and this whole problem (and why it's basically impossible if you allow much variety, much like Goedel's sufficiently powerful number system Incompleteness Theorem) in Metamagical Themas."
Having a sporadic section has worked out well with our recently introduced MailBITS, so we're introducing another section, called TechnoBITS. Here you'll find little bits of information about new and emerging technologies (real ones, this week) that don't warrant a whole article.
Intel recently showed a prototype 100 MHz version of its 80486 chip at the International Solid State Circuits Conference. The completed chip will probably be slightly less than three times as fast as the 33 MHz version. The prototype chip works with 5 V of power at room temperature with a normal heat sink, but Intel needs to iron out problems such as RFI noise and locating other chips with speeds fast enough to complement the 100 MHz 486.
MIT and IBM are working on a chip that will be far faster than any current chips if they can make it work at normal temperatures (my feet get a little chilly when the room temperature hovers around absolute zero). The chip doesn't use a stream of electrons like current chips - instead it turns on and off a single electron, which is far more efficient.
I love new input devices, and BioControl Systems of Palo Alto may have one of the best so far. It's a device mounted on a headband that monitors the electrical field movement of your eyes and moves an object on the screen accordingly. BioControl Systems is looking for capital to go beyond the current prototype, possibly first into video game control, but eventually into mouse-type manipulations. It sounds like a wonderful idea, but could play havoc with your eyes after a while.
MIPS and National Semiconductor have each come up with 64-bit RISC chips that are highly desired by high-end graphics people. A 64-bit chip will help in two areas - number of colors available and addressable memory size. Considering that the current 32-bit chips can have 24-bit color and use the additional 8-bits as an alpha channel for transparency (as the NeXT does), a 64-bit chip could have 48-bits of color and 16-bits for an alpha channel. By my rough calculation, that makes for 2.814749767 E14 colors. I'm no expert, but I doubt that any monitor made could display that kind of subtlety, and the human eye might not be able to distinguish it even if the monitor was capable of it. The other problem is that a 32-bit chip can only address 4 gigabytes of memory, which is apparently starting to cramp some image processing people. A 64-bit chip, though, is edging into the whomptillion (a unit of measure usually used with respect to the US federal deficit) range - 1.844674407 E19 bytes of memory. I couldn't find a real term for such a large number, but either 18 megaterabytes or 18 gigagigabytes should work. Take your pick One way or another, it should hold these people for a while.
PC WEEK -- 21-Jan-91, Vol. 5, #3, pg. 1
BYTE -- Apr-91, pg. 27
InfoWorld -- 25-Feb-91, Vol. 13, #8, pg. 22
New York Times -- 17-Feb-91, pg 8 (Business section)
BYTE -- Mar-91, pg. 32
InfoWorld -- 25-Feb-91, Vol. 13, #8, pg. 22
Double Helix has the honor of being one of the first and most popular Macintosh database packages. The program has had many changes over the years, few of which I've seen, since I started working with the program last summer. At the time, Double Helix impressed me because of the ease with it allowed me to set up a relatively complex database, though I later discovered that it allows you to set up the same structure in several different "right" ways. Needless to say, some ways end up more correct than others, or in my case, faster than others after 9000 records go in. Double Helix also suffered from not supporting the Macintosh interface completely despite its graphical programming environment.
Version 3.5, which Odesta announced at Macworld Expo in Boston last August, will fix several of these shortcomings this spring. Odesta has announced that it will include Double Helix's multi-user capability (along with a license for three users) with every copy of Double Helix 3.5. Previously you had to buy the multi-user version separately, which was a pain, although it was trivial to convert a database to multi-user use.
Double Helix was never slow, but it required some tricks to maintain a fast pace. Members of the Double Helix SIG on America Online provided invaluable help to me when I was struggling with the tricks and techniques. Their feedback to Odesta has undoubtedly helped, what with version 3.5's speed increases. Odesta claims that some functions have been speeded up by as much as an impressive 800 times. Odesta added multithreading, which allows Double Helix to execute queries while handling data entry. It runs faster over a network, a capability which should help Double Helix compete with other multi-user database programs, since it is designed for a client-server environment, unlike the other relational databases out there (although 4D does have a server now, I guess).
The most obvious enhancements to Double Helix are in the new objects it supports, check boxes, radio buttons, and pop-up menus. There were a number of tricks for simulating all three with Double Helix's old tools, but the tricks were without exception clumsy and slow. In some ways, I wish I could redo my database with version 3.5 because a number of screens would be easier to use (and much nicer looking as well).
The final enhancements that will please Helix developers include the ability to shrink a database, reclaiming unused space, and a Get Info... tool that lists the relationships between objects. The shrinking tool prevents the database from becoming too large in the course of development and testing, as currently happens. New records use up wasted space in the current version, but when you've got a nine megabyte database, it would be nice to dump all the data to a file, reclaim the space, work on the small database, and then load all the data back in when you're done. The Get Info... tool removes a great deal of frustration when you want to delete an icon that is being used by another icon somewhere else. Double Helix prevents you from deleting icons that are in use, but there used to provide no way to tell where it was in use. I spent hours on occasion, searching for the relationship that prevented me from trashing an unused icon.
Of course, such wonders as these come only for a price, and the list price has risen to $695. You will receive a free upgrade if you subscribe to Odesta's Tech Connect program, otherwise, it costs $99.
Odesta -- 800/323-5423 x 234
Sundry kind souls on the AOL Double Helix SIG
MacWEEK -- 02-Apr-91, Vol. 5, #13, pg. 5
MacWEEK -- 12-Feb-91, Vol. 5, #6, pg. 6
InfoWorld -- 11-Feb-91, Vol. 13, #6, pg. 40
Insignia Solutions is not sitting still with its SoftPC emulation software. Earlier this month, Insignia began shipping a new version of SoftPC tailored for use with the older and less powerful Macs, the Plus, Classic, SE, Portable, and LC. This version of SoftPC emulates the 80286, as does the older version with the EGA/AT Option module, but the new version does not include support for EGA graphics, expanded memory, or for the math coprocessor (which isn't surprising since none of those Macs have a coprocessor). The new version is cheaper as well, at $199 list, but keep in mind that it will not run on an SE/30 or Mac II-class machine. For that you still have to get the $399 SoftPC 1.4 and optionally, the $199 EGA/AT Option. Insignia may combine SoftPC 1.4 and the EGA/AT Option, but there's no telling when. If you're curious about SoftPC, hold on a bit, we're working on a review of it.
Of course, not everyone who wants to emulate a PC has a Mac. For those of you with NeXT machines (out of curiousity, do any of you have NeXT machines?) there's now SoftPC 2.0 for NeXT workstation, which is similar to SoftPC on the Mac. There are two exceptions, which point toward coming attractions in the Mac versions. First, SoftPC for the NeXT runs at about the same speed as a 12 MHz 80286 machine (SoftPC on my SE/30 is only about twice the speed of an 8086-based XT, quite a bit slower than on the NeXT), and second, SoftPC on the NeXT runs programs that require VGA graphics, but cannot display all of VGA's colors, mapping them to EGA's colors instead. Of course the SoftPC screen is a 640 x 480 window inside the massive NeXT screen, and if you were sufficiently masochistic, you could run Windows in that window.
VGA graphics are in hot demand on Usenet as well, and a number of people said that they had heard that Insignia is working on a Mac version of SoftPC that emulates the 386 and VGA. A version of SoftPC with these abilities might be available sometime in the middle of this year. I hope that if Insignia gets the VGA emulation working that they also speed up SoftPC a bit, since VGA graphics applications for the PC could be painfully slow under SoftPC.
Insignia Solutions -- 408/522-7600 (0494 459426 in the UK)
Matthew Kendall Howard -- email@example.com
Todd A. Green -- firstname.lastname@example.org
Andrew Theodore Laurence -- email@example.com
Jonathan Thoma Sweet -- firstname.lastname@example.org
MacWEEK -- 12-Mar-91, Vol. 5, #10, pg. 5
InfoWorld -- 18-Mar-91, Vol. 13, #11, pg. 8
PC WEEK -- 25-Mar-91, Vol. 8, #12, pg. 36
Yup, we made up almost the entire article (other than the bit on SchoolTalk - can anyone give us more information on that?) last week on SentientNET. Nothing in our April Fools Day issue is impossible and a lot of it would probably be a good idea. However, SentientNET, unlike the subjects of the other articles, does exist in a slightly different form. The Open Software Foundation's (OSF) Distributed Computing Environment (DCE) provides exactly that sort of virtual CPU power over LANs and WANs. (The abbreviations are coming hot and heavy now, with OSF's DCE on LANs and WANs.)
At CeBIT in Hannover, Germany, OSF showed an off-the-shelf application, Market Minder, running under OSF's Motif interface on machines from each of five manufacturers - IBM, DEC, HP, Groupe Bull, and Siemens-Nixdorf. The program filtered data from the New York Stock Exchange, working with the consolidated processing power of all the machines.
This sort of thing is extremely helpful, because it allows most types of computers to work on problems that would otherwise be too processor-intensive. It means that less-capable computers will stick around longer because they will still have some use, though with the maintenance costs on some of these mainframes, it may still be worth recycling them. Lots of hardware and software manufacturers have licensed DCE, although I didn't see Apple's name in the list (anyone know about this?). They'd be stupid not to at this point.
The hardware included an IBM mainframe running MVS, an IBM R/6000 running AIX, a IBM PS/2 running OS/2, and a VAX under VMS. The network used was Ethernet. I don't know for sure, but I wouldn't be surprised if a Macintosh running A/UX and Motif could also work with DCE. As far as the software goes, DCE is a layer of network software that lives between the OS and the application, distributing processor requests appropriately. DCE is smart about what it distributes where, so it wouldn't assign a section of a program that relied heavily on numeric calculation to a machine without a math coprocessor if it could avoid doing so. DCE is designed so users should never notice a difference different, short of tasks taking less time.
OSF -- 617/621-8700
Open Software Foundation propaganda -- email@example.com
InfoWorld -- 18-Mar-91, Vol. 13, #11, pg. 8
Apple has begun to step down from its ivory tower, or perhaps it's being pulled down by gravitational market forces. As much as the Classic is selling like hotcakes (pretty soon you'll be able to buy Classics in department stores and roadside diners - I'm not entirely kidding on that first one), Apple still has a ways to go before it's installed base can compare with the base of Windows-capable machines. Apple has been working on graphical environments for a number of years and considers itself an expert. Why not expand into the Windows market, and why not use Claris to do it?
Claris is working on applications for Windows, and although I'm not sure of what they are, I wouldn't be surprised to see Windows versions of MacWrite II, MacDraw Pro, FileMaker Pro, and the as-yet-unnamed spreadsheet with technology from Wingz in it. MacWrite II has a decent chance, although Word for Windows and Ami Professional are pretty good word processors. On the other hand, if MacDraw came in at a relatively low price, it might be able to make off with a good bit of the market from higher end programs like Illustrator and Corel Draw. FileMaker also stands to do well in the Windows market and wouldn't even need a name change like MacWrite and MacDraw would. A spreadsheet will be hard put to compete with Microsoft's nicely done Excel 3.0.
I'd like to see Claris release a Windows version of HyperCard that could run all Mac HyperCard stacks. If Claris distributed it cheaply, as it does with the current version (cheap or free unless you want the documentation), and sold it to developers for the $49 price, it would be an instant and complete hit. ToolBook, HyperPad, and Plus have generated interest in the area, but none of the three has taken over, whether it be because of speed problems, lack of graphics, or general flakiness. A speedy, stable version of HyperCard that could directly read all current stacks would be ideal. The other application that would endear the Windows market to Apple would be an enhanced version of the Finder (I say enhanced so it could deal with program groups and the like) to replace the brain-damaged File Manager and Program Manager. If Apple/Claris has the experience to provide these programs, I say, "Use it!"
Another interesting thrust out of the Mac market is with DAL, Apple's Data Access Language, which is a superset of SQL (Structured Query Language, pronounced 'sequel,' for some reason). Apple just licensed DAL to a couple of third-party vendors, including Blyth (makers of Omnis 5), Novell, and Data General. DAL integrates data access over many different platforms, including DOS and Windows, something that large companies with lots of data in lots of places like. It's definitely not aimed at normal people like you and me, but hey, if it helps insert the Mac further into corporate America, so be it.
Writing for the Windows market would cause some interesting conflicts, though. Apple has its own handwriting-recognition extensions for the Mac which it regards highly but is still considering licensing PenPoint from GO to help stem the force of Microsoft's Windows handwriting extensions. If Apple teams up with GO and IBM (remember IBM, GO's main partner?) it would be a potent combination against Microsoft. Apple might integrate support for PenPoint into the MacOS, not to use PenPoint's handwriting recognition since Apple thinks its version is better, but to increase compatibility with notebooks using PenPoint. That would increase the appeal of PenPoint to third-party hardware manufacturers and would help slow Windows/H or whatever it's called. Apple and/or General Magic might use PenPoint directly, which would mean that machines from Apple and IBM would for the first time be completely compatible. Scary thought, but it's rumoured that General Magic might have something out in a couple of months. I'm looking forward to it, whatever it is.
MacWEEK -- 26-Mar-91, Vol. 5, #12, pg. 1
InfoWorld -- 01-Apr-91, Vol. 13, #13, pg. 1, 8
Non-profit, non-commercial publications and Web sites may reprint or link to articles if full credit is given. Others please contact us. We do not guarantee accuracy of articles. Caveat lector. Publication, product, and company names may be registered trademarks of their companies. TidBITS ISSN 1090-7017.