|This article is of interest to the following WikiProjects:|
- 1 Service and Inventory Issues
- 2 NCR's logo has just changed : http://www.ncr.com/images/site/affiliate/ncrLogo.gif (i don't know where to find a good quality one, but if anyone could help on this)
- 3 Missing a LOT of important product information
- 4 Why AT&T bought NCR
- 5 First mass storage
- 6 Template
Service and Inventory Issues
NCR is not my area, but they do have a very large service business. There was also a major issue with the NCR inventory system that caused major havoc some years ago. --Gadget850 13:53, 20 October 2005 (UTC)
NCR's logo has just changed : http://www.ncr.com/images/site/affiliate/ncrLogo.gif (i don't know where to find a good quality one, but if anyone could help on this)
Rama, Does anyone have any info about NCR Unix (AKA MP-RAS)? I didn't see any mention of it on this page? MikeDawg 16:56, 12 October 2006 (UTC)
There is nothing here about NCR's manufacture of carbonless carbon paper which, until 1970, contained PCB's and resulted in widespread environmental contamination. Sanjour 20:07, 21 June 2007 (UTC) William Sanjour
I'm not real proficient in doing edits but for the record the logo and associated branding have now changed (to solid green). Someone more proficient than me may want ot change it.RedRiverGorge (talk) 22:02, 3 December 2008 (UTC)
Missing a LOT of important product information
I worked for NCR Corporation as a Field Engineer in what might be consider its heyday, the 1970s, prior to its being acquired by AT&T, when the entire Point Of Sale field devolved into cheap PC-based systems. [[ I was glad to see this write-up, it mirrors my own timeframe and experiences with some of the leading edge NCR technologies of the early 70s. Nice work! I've added a few comments here and there for clarification, hope you don't mind DrZeuss qlxpnzy (talk) 01:27, 30 August 2009 (UTC) ]]
My main area of expertise was their mechanical and electronic cash registers, and I worked on just about everything that was in use at the time. None of the products I worked on, or the products my peers worked on, are even mentioned in the above article.
Here are some worthy of note, in roughly chonological order.
The work "class" is used because the various models could be configured in various ways depending on the intendede use. For example, a department store model was very different from a restaurant model.
Mechanical cash registers:
100 Class -- probably the last of the manual, uncranked cash registers, powered by the keys themselves. Sometimes seen in movies. Direct descendent of the fancy antiques.
Class 21,22 and 24 -- electric, very popular, usually seen in restaurants, commonly seen in movies.
Class 51, 52, 53 -- popular in department stores and businesses needing more functionality than the 21/22/24 series. The 53 had a credit card reader and produced a computer-read printed journal paper tape.
Class 5 -- probably the most sophisticated and complex mechanical cash register ever made. Programmed with sliding metal plates. The design could easily have been modified to make it a general-purpose, if limited, mechanical computer in the vein of Babbage's Difference Engine. I've alwasy felt they deserved more recognition in the history of computing.
The class 5 was apparently related to a Class 3, which I've never seen, and which was apparently used in Europe. I don't believe NCR produced any new mechanical cash registers after the Class 5.
Electronic Cash Registers:
Some of the early ones were very sensitive to electrical noise and static electricity. I never could figure out what they used as CPUs, but they preceeded personal computers and showed no similarity to them. [[ NCR designed and manufactured a variety of proprietary LSI chips and these CPUs were likely a MOS version of the new 4004 CPUs (4-bit nibble) Intel made popular. DrZeuss qlxpnzy (talk) 01:27, 30 August 2009 (UTC)]] Plugging a bar blender into the same outlet as the bar cash register cause very unreliable operation. In the early days, non-volatile RAM chips were not available, and expensive magnetic core memory or large internal batteries were used to retain data. In most cases they had 1KB to 4KB of RAM, all of it non-volatile.
Class 220 -- freestanding, common in small retail outlets, had big lead-acid Gell Cell batteries to hold data in case of power faillure.
Class 250 -- a powerful register that replaced the Class 5 in clothing stores, bars, etc.
Class 255 -- used similar components to the 250 and 280, but with little or no autonomy. It was completely dependent on a back-office 726 computer which used the 605 processor (see below). [[ There was a reduced functionality version of the 255/726 marriage for small stores with only one or two on-site POS 255s (think "drug store") that stored the day's transaction activity on a live cassette tape. One terminal was then set up in "host mode" and awaited a "call" to upload the cassette tape contents to the host each night. DrZeuss qlxpnzy (talk) 01:27, 30 August 2009 (UTC) ]]
Class 280 -- the most common department store computer. Worked semi-autonomously, communicating via 40kbps direct current terminal-demand connection to store level concentrators which linked to 725 controller and from there to a NCR Century mainframe.
Class 285 -- a variation on the 280 used in banks, credit unions, and the like.
I did't work on these, so can't speak knowledgeably about them.
605 -- a 16-bit processor used in the Class 399 accounting system, the 721 and 723 communication controllers, the 725 department store level controller, the 726 retail controller, and others, including several of NCR's general-purpose minicomputers. The CPU alone comprised of four TTL circuit boards, each about 10 inches by 14 inches (as I recall), plus separate circuit boards for memory, communnications, etc. The clock cycle was 400 nanoseconds, and most instructions executed in one or two clock cycles. The instruction set was optimized for communication, and was somewhat comparable to Motorola's 68000 series chips. I'm biased, but I always thought it was well-designed and deserved more recognition. [[ Strongly agree with your opinion. You might want to add, for clarification, that the four boards that composed the 605's CPU were discreet TTL component versions of what's found in today's single-chip CPU. Nothing in the MOS LSI arena during that timeframe would execute as fast as its TTL brethren. The 605s were speedy! DrZeuss qlxpnzy (talk) 01:27, 30 August 2009 (UTC) ]]
The 605-based minis stick in my mind as being the first time I ever played a computerized version of Star Trek, a game played using "ASCII graphics" on 24x80 text terminals. Not sure where it originated. It eventually showed up on DOS-based PCs. Also, NCR's IMOS Cobol manual was the best I ever encountered.
Century series -- I know little about these, but I believe they competed with some of IBM's smaller systems.
Some general comments:
The previous reference to NCR paper (marketed as No Carbon Required) needs detail. NCR paper was developed by NCR based on their development of micro-encapsulation technology. This technology was a key to a number of other uses, including timed-release medication, etc. At the time it was developed, I don't believe the risks of PCBs were known.
NCR played a part in the Wright Brothers' success, helping with their construction of wind tunnels and so on. The naming of the Wright-Patterson Air Force Base backs this up, and there are examples to be seen in the Dayton air museum.
126.96.36.199 18:37, 8 October 2007 (UTC)
Why AT&T bought NCR
I'm not sure how to incorporate this into the article, but one of the main reasons that AT&T bought NCR was that AT&T desperately needed a sales force.
Before AT&T was split up in 1984, it had a captive market for its products, the local telephone companies, which were all wholly-owned subsidiaries. Thus, its marketing strategy was what has sometimes been described as the Field of Dreams scheme: "If you build it, they will come". After the divestiture, the telcos could buy from anyone and did. Northern Telecom, for example, had a telephone operator console that was almost as good as AT&T's OSPS at about half the price, so the telcos bought from Northern Telcom.
The leadership at AT&T realized that they needed to get a sales force really quickly, so they snapped up NCR in a hostile takeover. They figured that NCR's product line was not that dissimilar to their own (especially since both companies made minicomputers). What they didn't realize was that the NCR company culture was so different from AT&Ts that integrating the two was extremely difficult, and many of the NCR people resented AT&T. It was not a marriage made in heaven. Jhobson1 13:26, 22 October 2007 (UTC)
First mass storage
The article currently says, "In 1962, NCR introduced the NCR-315 Electronic Data Processing System which included the CRAM storage device, the first mass storage alternative to magnetic tape." I've figured out what you mean - automated storage for non-volatile data not used in any current job; comparable devices were automated mag tape reel or mag tape cartridge libraries, e.g. IBM 3850. The problems with the current phrasing are: mass storage takes a different view, including disks as well as tapes and thus giving the impression of referring to something smaller whose contents would all be online at the same time; "... alternative to magnetic tape ..." in the current article gives the same impression. I'm changing the phrasing to "the first automated mass storage alternative to magnetic tape libraries accesed manually by computer operators." Philcha (talk) 12:27, 25 November 2007 (UTC)
The links about cuts in 2007 are broken.
There are many dead links here, so here comes some more general perspectives on how Wikipedia can use references.
I cannot but wonder what will happen when, after some time and lots of added knowledge, most of the references at the bottom of wikipedia pages become obsolete.
Should there be a way to copy link+reference content into some kind of article-database, provided that e.g. Daily News accepts this usage of their site? Could we arrange with the "wayback-machine" some agreement that when one of the better wikipedia pages points to an article be it in a newspaper or a university site, then this link-target can be sent to the "wayback-machine" which then copies the text (mostly text references are needed.)