Steve Jobs's legacy

[Omaroo (Chris Malikoff)]

It was still Apple who convinced Warnock to adapt their PostScript raster processor to drive laser printers, and render vector art and typefaces in difference to other other faster bit-map print engines at the time. The LaserWriter's internal RIP required more processing power than the Macs they were connected to in order to output anything at near what could be considered reasonable speed.

I agree that LocalTalk was the slow part, and many of our customers specified EtherTalk as their preferred physical layer and ran NetBEUI, DECnet and IPX - it was a right-royal mish-mash until TCP sorted it out later. Then there was TokenRing, or "TokenTalk".... Golly they were great days Bring back SDLC I say!

[gary]

Originally Posted by Omaroo

What you say is true Gary - except that in "commercial" I was, I apologise, intimating "home" computer - not workstation. Apple did, indeed, bring this technology to the masses where it was previously so far out of reach it was laughable.

Hi Chris,

Thanks for the clarification but no need to apologize.

I also unpacked the very, very first IBM PC that landed in Australia - and we engineers, too, laughed at it saying that "it'll never catch on"

Now there's a claim to fame!

But I know exactly what you mean. The university VLSI research group I was
employed by had received some generous funding out of Yorktown Heights.
Hoping we might be able to use it, IBM also decided to ship us one of their
latest offerings - an IBM PC.

The research group had its own VAX 780 that only a dozen of us shared,
mainly through VT100 style terminals but also by way of a couple of AED
graphics terminals with large color displays and tablets with pucks rather than
a mouse in those days. So a lot of the full custom integrated circuit design was
done on these big graphics terminals connected to the VAX housed in an
airconditioned machine room and if the design layout work became really intensive,
the call would go out for everyone else to go have a coffee break so the graphics
work could continue in real time.

So this comparatively little baby PC turns up and we all scratch our heads as to
what we could use it for and we all have a chuckle. It sat on one of the lab
workbenches outside my office and now and then you might catch someone
who should have been working on their thesis or whatever playing some game.

But of course the personal computer did eventually come of age.
And some of the real champions of course are those who you never hear their names.
Inside the various products from various companies mentioned in this thread
are devices and software designed by large teams of engineers. The true success
of many of these products relies on the very high levels of integration within the
chips and their packaging. Semiconductor engineers year after year have
managed to make geometries smaller and smaller, fitting more and more
functionality onto a device whilst using less and less power but running at faster and
faster speeds.

Some of these engineers will be veterans, but many will be very young, fresh out
of university and only in their early twenties. The pace of development in the
semiconductor industry is so relentless that many will effectively burn-out
by their late twenties as coal-face designers and will take on senior chip architect or
management roles. With Non Refundable Engineering (NRE) costs on chips now
at multi-millions of dollars for the first prototype device - just one team member
makes one mistake and you kiss millions of dollars good-bye - the pressures
on these young engineers to deliver on-time and on-budget is enormous.

Many of the electronic products that will dazzle us a decade from now will
end up being designed by people who are currently only 12, 13, 14 years old and don't
even know they will be engineers yet.

[gary]

Originally Posted by Omaroo

The LaserWriter's internal RIP required more processing power than the Macs they were connected to in order to output anything at near what could be considered reasonable speed.

Hi Chris,

It was actually a little trickier than this. Without sufficient processing power
and given the limitations of the amount of available DRAM, you could run into
a situation we referred to as "page too complex" that would prevent you
from successfully outputting a page at all.

The LaserWriter's controller used a 12MHz Motorola 68000 and the LaserJet
an 8MHz Motorola 68000 and both used the same Canon CX Engine, but the
Apple LaserWriter sold for US$7000 but a HP LaserJet sold for half the price.

The primary difference was the amount of DRAM their controllers had. 128K in
the LaserJet but 1.5MB was required in the LaserWriter.

I designed a large number of laser printer controllers and I use to spend a large
amount of my waking hours and many of my sleeping ones thinking about
memory. Memory was very expensive. It dominated the cost of a laser printer controller.

The problem for the controller is that once you issue the print signal to the printer
and the motors start turning, you are committed and can't switch them off in
the middle of printing. You have to rasterize in real-time. And if you have less DRAM
available than what it takes to represent a 300dpi A4 bitmap, then you will need
to rely on a lot of processing grunt (either via processor or ASIC) to render a
band at a time down the page trying to keep just ahead of the print engine.

If on the other hand you had the luxury of enough DRAM to represent the entire
bitmap plus some for the data input buffers and some for the processors
stack and memory requirements, you had to worry less about real-time
constraints, but it made for a very expensive laser printer. Adobe used this later
approach and years later, everyone else did once the DRAM became more affordable.

But in that first decade, by designing laser printers that required less DRAM,
it meant companies like HP could sell them for less. This strategy, amongst others,
ensured HP got a stronger foothold in the market and thus allowed it to become the
world's number one printer manufacturer whereas others, such as Apple missed
the boat.

Once HP introduced scalable fonts, except in areas such as the graphics arts
market, CAD and a few other niche areas, few consumers required or for that matter
appreciated the differences between PostScript and PCL.

So in some ways it might be argued that Apple had a perfect opportunity to
monopolize on the laser printer experience out of PARC and to their credit they
chose, at least by way of PostScript, a technically very elegant solution.
But as a commercial, money in the bank, share price at the end of the day decision,
someone at Apple - perhaps it was Steve Jobs - made the wrong call.

[Omaroo (Chris Malikoff)]

Hey there Gary

It's certainly interesting hearing of your intimate connection with this technology.

I was involved with getting DTP off the ground here in Australia, primarily through various Apple houses such as Avanté Systems and Logical Solutions. We were only really interested in vector output to offer to the large magazine and newspaper publishers who were looking at divvying out previously-expensive proofing duties to cheap but capable laser printers - rather than clogging up their Atex and Agfa RIPs and imagesetters which were too busy producing bromides to go to plate. Plain paper mono vector proofs were suddenly acceptable where PCL really wasn't as it was device dependant. PostScript was seen as the real answer as it gave realistic comparisons with the big Agfa and Atex cousins as PostScript was designed to output to any PostScript device in exactly the same way. We needed apples for apples as it were.

HP LaserJets were heavily employed at the advertising houses for proof of concept and general design and desktop publishing work. We sold many, many of these at Avanté and Logical. We sold plenty of LaserWriters to News, Fairfax, ACP and Pacific Publications, as well as large agencies such as Mojo and Neville Jeffress - again for reliable, consistent plain paper proofs.

Admittedly, the call for PostScript in the general office or home was minimal. The cheaper printers sufficed there, and at that juncture still had a hell of a time removing dot matrix from the equation for general printing.