Oct 86 Mousehole
|Column Tag:||Mousehole Report
By Rusty Hodge, Mousehole BBS
More On Hard Disks
Rusty Hodge, Sys Op
The battle of the hard disks continues. A few of the newer players are the ProApp 20, the Easy Drive, the HyperDrive FX-20 and a few others I'm forgetting to mention. So what is so special about these drives?
The ProAPP will work on a SCSI port, or on a 512 through the disk port, or on an Apple //e or //c. You can partition the drive between Apple // and Mac, even use them simultaneously to a limited extent (if one is accessing the drive while the other tried, you'll get a nasty I/O error). They tell me it's not supposed to work that way, but it does. Speedwise, the ProAPP seems about as fast as the Dataframe, although the particular unit I have is a bit noisier.
The Easy Drive is the true mass storage winner. About as pretty as a LoDown (read: plain), but you can't beat 45mb for $1500 and 60mb for a couple hundred more. Can't say more because I only played with one briefly but hope to have one by next month (hint hint...).
Last, but we hope not least is the HyperDrive FX-20, General Computer's first entry into the SCSI market. Novel features include keeping the SCSI parameters in the drive's own parameter RAM, therefore allowing you to graphically set it up on the SCSI chain. It also comes with most of the traditional Hyper goodies including the backup and security programs, however no Drawers and no HyperNet! It is also big, boxy and ugly... but then who am I to say? (Still, a lot of people think the IBM PC is a great looking box...)
Did everyone know what happened to Tecmar? They were having some financial problems, but someone thought they'd be a good investment, so Tecmar got bought (believe it or not, Tecmar was privately held).
Anyway, three weeks ago, about 110 people were laid off, around 1/5 of the employees. The guy I report to is still there. We talked about their new product, but I can't say. I did get some information that MacDrive owners should be interested in. To print with the MacDrive, you need to use the Imagewriter driver that was supplied on the MacDrive 2.2 boot disk. Yes, this sucks. This means that if you have an Imagewriter II, you can't use any of its wonderful features. Sorry. Everything else in the 3.2 System works.
He told me that there WILL be a new version of the MacDrive software, version 2.3. It'll have some neat stuff and will fix the Imagewriter problem. One other thing. HFS works fine on a MacDrive, just make a single ten-meg volume. The Volume Manager doesn't know about HFS, so it'll say "This volume has not been initialized" but fear not. You can still mount the volume, and set it to be the startup volume. The only problem is that the System file needs to be in the Root. This in turn means that ALL system files need to be in the root, because the "System Folder" is only magic if it actually contains the System file.
There is a way to make the Tecmar run HFS and still keep the system stuff in the system folder. How do I know? I've got one doing that. I think it is because I'm using the HD-20 startup file. The 128k ROMS should do the same thing according to The Anarchist, but we haven't been able to try this.
A new book is on the shelves by the same company that produced Chernicoff's stuff called Object Oriented Programming on the Macintosh, by Kurt Schmucker. A quick scan showed a ton on MacApp. It's been over a year and a half since I spoke to Tesler about MacApp. There's a book on it, lots of press, but no MacApp! Reminds me of MacBASIC! [Note: Kurt Schmucker is doing a special issue in MacTutor on MacApp. Watch for it shortly. -Ed.] RE: the Object oriented programming course at NB. A bunch of us went to hear Mike Miller talk about development over a year ago, and he preached ExperLisp... well? He also told us the LaserWriter will never be a major product.... He also told us that C is a terrible language and nothing significant will ever be produced with it on the Mac...... Yawn.
Hard Disks at the Mac Expo
One thing that I noticed at the show. Almost everyone who was not promoting one brand of hard disk or another (in other words, who got to choose which hard disk to use) was using the DataFrame 20. There were a couple of exceptions, but the vast majority were SuperMac DataFrames. I think that this indicates something - what, I'm not sure. [Remember the first MacWorld Expo, where there were Corvus Omnidrives everywhere? Dèja Vu? -Rusty]
[Don't say that, Rusty, I just bought two Data Frames! Ouch. -Ed.]
I've been fooling around with the AST 2000 and 4000 and don't like it. First of all, it is VERY expensive. What do you get for the money? A hard disk system with integral tape back-up device. OK fine, it better knock my socks off. How does it work? Well, with the 4000 I want the ability to select a folder and backup everything in it. Sorry, no can do. Promised, but not ready. Oh well, how about a complete volume back up? Well, you can only back up 60mb, not the complete volume. How fast is it? I wish I could tell you this, but after waiting for a hour I gave up and rebooted. AST said, "You need the lastest version, it's faster and available on our BBS". OK, fine. Called them, downloaded the stuff, and installed it. Here we go again. "Complete (well almost a complete) Volume Backup". This time I rebooted after 30 minutes, never did get a backup. And another thing, what about the last 13 or so megabytes? Does it just get ignored, or can you back it up? The AST 2000 is a 20mb drive with a mini-tape drive. The tape device in the 2000 seems real bogus to me. Loud and slow. My biggest complaint against it is that you have to "Format the tape"! Why do you need to format tape? The format procedure takes over a hour; it doesn't come formatted from the factory. The first 2000 failed to format the tape. OK fine. "What's the problem?" I asked AST. They said I probably have a bad tape, try another one. I had been using the tape they supplied with the drive. The second tape failed, too. Two out of two, and no backup yet. Has anyone successfully backed up an AST? How many minutes per megabyte did it take? Our entire accounting software is stored on two 22mb drives. Including the time it takes to load in a fresh "unformatted" tape and push a button, we can back up 40mb in under five minutes. What is AST's problem? I like the speed of their 4000, but without a complete backup device, forget it.
The main reason you have to format the tape drive is that AST configures it as an on-line volume; that is, it comes up in the Finder with its icon and behaves like an HFS disk. To do a backup, you simply do a Finder select-and-drag to move folders and files onto the tape. Their low-level device handlers are supposed do the rest. I know the thing works, since I saw one up and running at the Apple Business Forum in Long Beach. The price, however, seems a bit steep. AST, like Iomega Corp., is a big player in the corporate MIS market. I would recommend their products to a large buyer, but steer the individual user to something less expensive. My recommendation for a SCSI device at this moment is the DataFrame 20 by SuperMac. It's fast and reliable. The only person I know who had a problem with one called the company and they shipped a replacement to him before they received his defective unit in return! There is no reason, with a SCSI interface, to have the tape backup unit integral with the hard disk. I would wait for a SCSI tape backup device from a respectable manufacturer and purchase it separately. The other way to go for business buyers is with the Apple HD20 and a sizeable RAM cache. Speed differences are minimal if they have enough memory to work with. Contrary to general belief, the RAM cache is very reliable with System 3.2/Finder 5.3. I routinely run a 256K cache along with Switcher as my normal "shell." The Finder gets loaded into slot 1 at boot time, and I rarely have to leave it. Of course, I have a Levco 2M board....
The tapes DO come formatted from the factory. The tape you describe sounds like it: 1) has an adjustment problem that came out because 3M changed their media and didn't bother to tell anyone about it, therefore requiring that all the tape drives in existing AST-2000s be readjusted, or 2) the power supply is out of adjustment. The second cause is very rare, but I won't say that shipping something has NEVER caused a problem. If anyone has any questions about AST's Apple products, feel free to call me at (714) 553-0340. That is AST's direct line to their Apple products group. Ask for Jim in Tech Support.
The Open Macintosh is a lot more than just Hardware. Consider these few points. If we use Open Macintosh to dissolve the functional barriers between micro, mini, and mainframe we can get on with the business of creating higher quality products and services. Open Macintosh must be able to run Systems software like MS-DOS and Unix. [I don't agree. -Ed] Unix's environment is hostile, but Open Macintosh could (should) make Unix just like a form of MAC-OS. This will allow users (US) greater freedom of thought, which we will need in the present and future work place if we are to be competitive with all the other powers of the world. Open Macintosh must use a bus like the VME. The fact that few retail sales persons know what to do with a VME bus hurts, but this won't stop those individuals who can design total business systems.
Upgraded 512s and the keyboard
My Monster + has no problem (thanks to DM!), but after trying an Apple upgraded + I find that Draw's control characters don't work correctly. I never bothered to try any thing else. If your thinking of upgrading your 512 to a plus, add the cost of the keybd upgrade, you are gonna need it! Boy are my dept's Bernoullis failing! Almost everyone of them is getting flaky. One worked okay after a cleaning, while another was totally trashed after a cleaning. I am sooooo glad I never got one on MY desk! Re AST: please continue posting on its performance. I'm probably going to be justifying at least one purchase this year and I'd like to know if it will go the way of Tecmar's MacDrive (yuck) and the printer port Bernoullis (razzzz). I've got an HD20 (yeah!) and after 6 months or so, not ONE hangup.
New Apple Product
I just spoke with a friend of mine who works for a major national account of Apple's, i.e. a company that buys LOTS AND LOTS of computers from Apple. They are also a major contractor, like Rockwell, but not quite... Anyway, his associates, not him, have SEEN the new Mac. Scheduled to debut in January it is supposed to retail for $4000. It will have slots, a separate, optional monitor, support up to 8 megs of RAM on the motherboard and run on a 16mHz 68020. It will, in other words, cook!! The impression seems to be that Apple will position it opposite IBM's AT. Also, it will come with an internal hard disk. (I assume a fan will be there too, though he didn't know). Also, according to various price lists that Apple releases to its dealers, Major Educational Institutions, and Developers, one can read between the lines. The Apple Daisy Wheel printer has been discounted for the last several months. Now it's gone all together. That implies a new product from Apple that will take its place. Also now missing on one or more price lists: the 10 mb profile, (though the 5mb is still available), the wide Imagewriter and the //c flat screen. Rumor in the trades is that a dozen products will be announced this September. Besides the //GS, there will quite probably be a new // and Mac compatible hard drive, (most likely a SCSI drive), and quite possibly new lower-cost laser printers. Anyone else heard anything? [Of course, they've probably all been introduced by the time you read this!- Rusty]
The New Mac?
I don't really know if the new Mac should even be called a MAC, since it also is rumored to run UNIX and MS-DOS and have the VERY POPULAR VME Bus structure for expansion. With over 20,000 VME bus devices that are already available, it can be just about anything you configure it for. Unfortunately, none of the retail salespeople that I know will know what to do with such a fabulous machine except sell it to run MacPaint and MacWrite.
Unfragment that disk
Hard disk slowing down? ALSoft Inc. has just released "Disk Express". This jewel claims to unfragment MFS and HFS disks by relocating files to contiguous blocks. It can purge old icons out of the desktop file and initializes sectors still containing data from deleted files (for security reasons). It's $30, call ALSoft at (713) 353-4090.
While some of you were fortunate enough to go to Boston, I made my way to SIGGRAPH in Dallas. For me, computer graphics is an all-consuming passion for the last several years; the Mac is one of the best parts of it. Most of us have seen incredibly real examples of glass with all of its reflective and refractive qualities, the phenomenal representation of Jupiter in "2010" and so forth. But every year is different, and better. Last year the neat effect was blurred motion (temporal anti-aliasing thanks to PIXAR), now this year's standard that a lot of the heavies use in their demos. This year, the theme was several new approaches to animation systems (why I attended) and the ability to render cloth incredibly naturally. We're definitely getting past the concept of key frames as a computer animation technique. Now there are very elegant script and hierarchical-based animation systems (kneebone connected to thighbone, etc., used to produce Jagger's HARD WOMAN video) (you should see what the Cray did to this in 70mm!), to dynamic and even behavioral and environmental systems. You can even create a flock of birds with the instincts of individual birds, and turn them loose, like the owl in the opening sequence of "Labyrinth". Mindblowing.
Dallas' SIGGRAPH was a busy place. A pretty good debate about window systems (Microsoft was there, what do you expect?) , a demo of "X" ...a distributed network environment that lets HP, SUN, VAX, IBM, and more talk to each other through windows over Ethernet. Neat. And Alan Kay gave a most impressive demonstration of the Vivarium project he's been working on. (Take above animation systems to create new interfaces, combine with the natural genius of children, shake well and watch out!)
As for hardware, last year it was 68020's, hot coprocessors, and the PIXAR. This year they had them cranked up to 25kHz, and they actually delivered some of Jobs' new machines. Several great new software packages for the heavies. From the AT end of things, I thought AT&T's TARGA board, and the TI 34010 were slick. I didn't think Old Blue's presence hung over the exhibition hall as much as it did last year.
What about the Mac? Well, there were hardly tny there! (Except in my hotel room, of course.) Usually they were attached to a high-speed laser printer or something. It's too bad MGM Station or EZDraft could show Autocad a thing or two. But the Mac's presence was definitely felt. I didn't know those recolored B&W movies were done on a Mac. I heard a bit about the Apollo/Mac system being set up. Lots of them were used in courses and panels and an incredible number of people said they like them and own them. A friend and I decided to initiate a SIG, and with less then one day's notice on a bulletin board had a good-sized audience, including one of SIGGRAPH's chairmen.
HINTS OF THINGS TO COME. Silicon Graphics, Apollo (and you can bet Sun, too) are working on boards for Johnathon. And face it, these folks know how to get power into image synthesis! Sun has also been working on a new windowing environment using Postscript. Levco is working on something with the 34010. Talked to Marc Canter at a party. He was "schmoozing" at full speed. He'll have a far much more powerful VideoWorks 2.0 out in February and some other neat stuff out soon. Dan Sadowski, the author of Comicworks, and several other Mac people were there checking things out.
There is a lot of anticipation here that with Johnathon, the Mac will start to play a very important role in computer graphics. With its anticipated power, open architecture and the almost industry-standard 68020, given the user base, it is an incredibly attractive package to the hardware and software developers here. Next year, SIGGRAPH will be at Anaheim July 27-31 - right in Mousehole territory. I really recommend it: five stars.
Rom Bugs <ugh>
I don't recall anyone mentioning this: but why does text in the editable fields in a dialog box move up and down after you have finished editing it? (I guess it only moves down: when you go to edit it again, it moves up). Has anyone else noticed this? The text edit bug is known at Apple, and since it's only cosmetic, they have no plans to fix it. [Apple plans to not fix a bug because it is "only cosmetic" Argh! Remember how we all liked the Mac because of the attention paid to all the little cosmetic details? -Rusty]
Things to watch out for in EXCEL:
When specifying a pathname in a macro, exceeding the Mac OS-allowed 63 characters results in an "Error in Macro" message. If you've specified a non-existent file (<63 char) the macro behaves normally and tells you it can't find the file and gives you the Standard File box. Microsoft customer service doesn't know about the 63-character limit (at least the ones I talked to). The "Bug of the Month" in MacUser last month doesn't fully explain the file deleting 'feature' of EXCEL. When saving a file, the old copy is deleted first. For example, when saving to a too-full disk, EXCEL will save as much as it can, then give the message "Disk is full" followed by "File Not Saved". What it doesn't say is "File Was Deleted"! You must save the file to another disk before closing or exiting, or the file will be lost. Another time-saving feature from your friends at Microsoft!
Here are some caveats about MacApp: MacApp provides for insulation between the programmer and the event loop. This may restrict your ability to do out-of-the-ordinary things. Some of the base routines in MacApp are far too slow. The simple text editor, for instance, cannot even keep up with s l o w typing.
My recommendation is to pass on ZBASIC for now. I spoke at some length with their technical rep at the Anaheim MACazine show, and he finally convinced me to buy the compiler. Since I truly enjoy writing quick applications in MS-BASIC, I figured that ZBASIC would be ideal: almost as easy and fully compiled... Well, first of all, the development environment is terrible. I know the company was trying to maintain compatibility with its existing products, but that's just death on a Macintosh. What you really have to work with is a line editor- pretty decent if you're writing code on a 64K CP/M system. It's useless for Mac programming. OK, I decided; I would use QUED and ZBASIC together in Switcher. This more or less allowed me to edit code, but then I tried to compile stuff. The first thing I tried was a sample terminal program contained in their manual. No workey. It turns out that you can't use ANY kind of event trapping (MENUS, DIALOGS, MOUSE, etc.) with the INKEY$ statement. This is because doing a MENU ON compiles a call to GetNextEvent before every line of your code. This removes all KeyDown events from the queue, so that INKEY$ never sees them. There is no way to read the damn keyboard on the fly while using menus or windows! When I reported this "feature" to Michael Gariepy, he was defensive and suggested it was a problem with my code. This went over real well. However, there may be a fix coming sometime soon. Dave Kelly recently spoke to the guy who actually did the coding for the Mac version. He relayed my suggestion to return keyboard events through a new DIALOG function, something like: C$ = CHR$(DIALOG(n)). The programmer said it would be easy to do, so it may show up in a future version. [Can someone tell us why they would sprinkle GetNextEvent calls all over the place in the first place? -Ed]
TML Pascal was the first Pascal Compiler that compiled to 68000 code, compatible with Lisa Pascal, that was available for the Macintosh. This, of course, has made it a very popular development system for those of us who prefer to program in Pascal on the Mac. Somehow, though, TML seems to have escaped any criticism. Don't get me wrong: I bought TML and have been using it to develop programs and desk accessories. The fact remains though, he compiler could do with a lot of improvement.
MY TML GRIPES [version 1.11]:
1) The compiler does not generate compact code. Example: num:= num + 1; is not compiled efficiently.
2) There is no support to call SANE directly.
3) The 'vanilla' mode leaves a lot to be desired. It leaves a vertical bar at the end of the text and does not provide support for backspacing.
4) When reporting errors, it does not report which procedure the error was in. In fact, it gives no indication whatsoever where the error occurred - it just informs you of the error.
5) The compiler is not 100% Lisa Pascal compatible.
6) There are not true Libraries, i.e. it cannot use the USES statement.
7) In 'vanilla' mode, it assumes too much. There should be the ability to use standard I/O in ANY window that the programmer sets up.
Also, automatic initialization in 'vanilla' mode of the Mac Traps should not be assumed. It should be an option.
There are a few other little items, but they've slipped my mind right now. TML may have good competition in the near future (Apple's MPW, Lightspeed Pascal, and Turbo Pascal). Then TML won't be the only Pascal compiler in the woods! The others coming out have some real good features. Let us be grateful for the first Pascal compiler on the market, but let's not worship it, as I have seen some Mac Developers do.
[I disagree with some of these items. The code generation is very efficient, more so than LightSpeed Pascal and pretty darn near 100% Lisa compatible, again, much more so than LightSpeed Pascal. Plus version 2 is rumored to run twice as fast as the compiler efficiency has been improved, and it supports both units and objects! -Ed]
Got my Lightspeed Pascal in the mail today.. After working with it for about 5 hours, I'm fairly impressed. It took a while to convert my TML code over. The code all worked fine, but my program was around 4000 lines long and I had to segment it up into 3 or 4 parts, since the LS doesn't want you to have more than 2000 lines in a Unit. The debugging is great! Like the MacPascal debuggers plus a assembly debugger called LightsBug which works very well.
It does have a few quirks though. A record in which I stored data in TML returned a size of 5336 bytes, but the same record on LS gave me 5436 bytes. Where is the missing 100 bytes? I can't figure it out, but it sure did screw up my Save/Open options that were reading 5336 byte files and giving Eof errors. One other thing I haven't figured out yet: when I go to my About box routine, I set up a window with NewWindow, then wait for a Mousedown event to exit. It works fine in the compiler/debugging mode, but when I compile it to an application it gives me a system error ID=11 after you click to exit the about box. Any answers? [Check the compiler option that controls who does the quickdraw and window initialization, you or LightSpeed. Try $I- option. -Ed]
Data Frame Drives
Had a talk with Steve Edelman, President of SuperMac Technology, about their Data Frame 20 hard disk. It seems that Walt at Coast Computer has returned his store's entire inventory of Data Frame drives and put a hold on further orders until he could get units with Seagate drives in them. He said he was getting too high a return rate on customer delivered units with the MicroSci drive. Steve Edelman said they use three different drives, the MicroSci, Seagate and La Pine drive, the last one being a newcomer. The drive manufactures claim their drives give a 1-3% failure rate, but surverys of store managers claim its more like 10 to 15%. And this isn't unique to SuperMac or any other vender, but a problem across the board. The drive makers just don't deliver consistant quality control from one batch of drives to another. Steve said they do extensive QA and burn-in to weed out any bad drives but still problems tend to come in cycles. He also clarified the OMTI controller board situtation. It seems from May to July, a terminating resistor pack was changed by one of the vendors that didn't make good contact with the motherboard, leaving the bus unterminated. The vendor has been replaced, the OMTI boards are fine now and they've seen no problems since then. Several other hard disk products also use the OMTI interface. So, I don't know why Walt is down on the unit, but after talking to Steve, I decided to lay down my money and takes me chances!
Oh yes, he also explained what the XP is all about. What they do is provide a daughter board that plugs into the Mac Rom slots, and you plug the Roms into the daughter board. On the board is a small Rom with a custom SCSI driver. An INIT resource changes the pointers to the custom SCSI driver so the XP drive runs at twice the normal SCSI port speed. Since each SCSI device uses its own driver, this does not effect other chained devices. Any data frame can be upgraded by buying this daughter board and inserting it in the Mac. The two drive units are essentially the same. Only the Mac SCSI driver software has been speeded up.