Why was the QL never successful?

[dex]

There were more PAL QLs produced, than the PAL version of Amiga 1200.
I.e., in Europe, we can say, Amiga 1200 was less successful, than the QL!

The QL was unsuccessful, because there was written in magazines, that it is not a success. This has big influence on common people's opinion.

According to the "Microcomputer case study" (https://sites.google.com/site/ql68kos/home/about-ql), Amstrad stopped QL in the moment, when it was well debugged, with increasing amount of software, with support of third party hardware producers, QL production was increasing thanks to the Samsung production plant, more and more people buying it (incomes from QL sales saved Sinclair temporarily from the Maxwell overtake).

[Peter]

To oversimplify it:

1. One 3.5" floppy instead of two microdrives.
2. 16 colors instead of 8 + blink, and more business-style 4 color scheme

Floppy had made the machine a little more expensive, but would have freed development resources. Floppy solutions were ready to buy, and prices falling. Time which went into microdrive would have been there to make the rest of the design more mature. And regardless the pros and cons, floppy was simply what people wanted.

In this case, the internal expansion slot would not always have to solve the floppy problem, but be available for RAM and other interfaces. Neater solutions, less boards sticking out of the QL side.

I do not see initial lack of mouse based GUI as essential, because the core system was already windowing, and in case of a large commercial success, the upcoming solutions like MICE and PE would have gained ground faster than thy did. Apple and Atari also needed time to develop. And compared to the PC, the QL was more than half a decade ahead, anyway.

8 colors plus hardware blink were pure nonsense, 16 colors coming at no additional cost, even freeing some logic space.

I do not agree to wider busses, higher graphic resolutions, more RAM etc. This would have added too much complexity and cost those days. The case design looked good, so no external keyboard.

[1024MAK]

I agree with most of what Peter said. However, as all this is written in hindsight, it's easy to work things out now

I think there are three things that would have vastly improved QL sales:-

• For serious business use you need a good keyboard, and as well as working well, it has to look good. If Sinclair had produced a computer with the type of keyboard that Memotech fitted to their MTX computer...
• As Peter and others have said, floppy drives were dropping in price. A single 3.5 inch drive instead of microdrives would have been better for the intended market.
• Unlike "home" computers, computers aimed at professional and businesses have to be seen as reliable, which means (as far as possible) stable and bug free. I understand the reason why Sinclair launched the QL when he did, but the company should have been up front with its customers and told them that delivery would be delayed. Then hold delivery until the machine and the ROM based OS and BASIC are ready. Only then send out to reviewers and customers.

In addition, some minor points that would have helped:-

• Instead of non-standard connectors, use D type connectors for serial and joystick ports.
• Instead of two crippled RS232 serial ports, one fully functional serial port and one "Centronics" parallel printer port (an IDC connector could have been used).
• A battery backed working RTC.
• The above changes would have increased the costs, so delete one of the joystick ports. Even on the Spectrum, there are not many games where two joysticks are used.
• A standard video signal so that existing designs of VDU ( monitors) could be used rather than people having to buy special QL compatible VDUs.

If Sinclair had done all this, the QL would have given its competitors a bit of a bashing.

Mark

[Mr_Navigator]

Also if he had sold 5 million units it would have made the QL succesful, but he didnt lol

Ihave to disagree with one specific point of Brane2

forget about ROMs and Superbasic. Implement just kernel ( basic multitasking stuff) in small ROM, with everything else that loads externally, but this small kernel thoroughly and with plenty of vision for growth and extension.

Not without SuperBASIC, this was a key thing for me and still is.

[Peter]

However, as all this is written in hindsight, it's easy to work things out now

As always.

For serious business use you need a good keyboard, and as well as working well, it has to look good.

Yes, I just didn't mention it, because the looks are good, just the "mechanics" needed improvement. Some delay in improving the keyboard would have been acceptable. I see it as something to work on while the first generation machines are already sold. Keyboard was an important cost factor, so simply buying a better solution could not have solved it.

Instead of non-standard connectors, use D type connectors for serial and joystick ports.

Definitely. This was done in non-UK version anyway.

Instead of two crippled RS232 serial ports, one fully functional serial port and one "Centronics" parallel printer port (an IDC connector could have been used).

Not sure, because one has to cut costs and development time somewhere. PAR could easily be covered by add-ons, especially as floppy would have been included, leaving more options for use of the extension slot.

A battery backed working RTC.

Maybe possible if some time was saved by not developing Microdrives.
A battery connector of some sort would have been sufficient.

A standard video signal so that existing designs of VDU ( monitors) could be used rather than people having to buy special QL compatible VDUs.

Maybe not, because 512 horizontal pixels were very important for decent text processing (80 characters with at least 6 pixels plus some frame pixels). Timings were very tight on the hardware side, and a solution so difficult that it probably did not justify costs and loss of time. There were non QL specific monitors which could handle the signals, and with a success of the machine, that kind of monitors would have become more common.

[tofro]

A battery backed working RTC.

Maybe possible if some time was saved by not developing Microdrives.
A battery connector of some sort would have been sufficient.

Well, it was there (and stil is...), but just never worked.....

A standard video signal so that existing designs of VDU ( monitors) could be used rather than people having to buy special QL compatible VDUs.

Maybe not, because 512 horizontal pixels were very important for decent text processing (80 characters with at least 6 pixels plus some frame pixels). Timings were very tight on the hardware side, and a solution so difficult that it probably did not justify costs and loss of time. There were non QL specific monitors which could handle the signals, and with a success of the machine, that kind of monitors would have become more common.

We should all be aware that there was no such thing as a "standard" video signal for a memory-mapped graphical display supporting >80x25 text. VGA was not developed back then - the closest were CGA (which was really crap) and (a bit after the QL) EGA modes - Even with EGA you had big problems finding a suitable monitor these days.

Tobias

[1024MAK]

To save any confusion, by standard video signal, I mean a 625 line, 15625Hz horizontal, 50Hz vertical video signal. Yes, it would only be any good if the CRT monitor used had a suitable small pixel size (high resolution) and the connection was RGB or monochrome composite video.

The Acorn BBC B computer can do it. So I do not see why the Sinclair engineers could not do it. And suitable monitors were available.

Remember, it is the cost of the complete system that a business would have been looking at. So using cheaper monitors would have been in Sinclair's best interests.

[1024MAK]

I don't see a "Centronics" parallel printer port as being hard to design or that expensive to include compared to a serial port. One 74LS series latch chip, a CS output from one of the ULA chips (to control the 74LS series latch chip) and using two I/O pins of the 8049 IPC (strobe and busy lines), plus of course the connector and your done

Now your "business" buyer can use nearly any parallel port printer on the market instead of having to buy a serial to parallel converter, or be limited to the more expensive serial printers.

Mark

[RWAP]

From memory, centronics printers at the time were not that common and there were at least a couple of different standards - serial printers were cheaper and more common (some of the printers from the late 80s had options where you bought a separate serial or centronics interface and plugged it in).

[1024MAK]

Printers must be unloved, as there is not much info on the web about launch dates of dot matrix printers (at least not so far, but I have not yet looked at magazine adverts).

http://en.wikipedia.org/wiki/Dot_matrix_printer
http://en.wikipedia.org/w/index.php?title=Parallel_port

Ahh, "the FX series was introduced in 1983". Link to source.
FX-80 review

Mark