Talk:Cullinet

Early comments

This article needs work. There are some minor problems regarding facts about the software mentioned itself. Spaceyankee 17:41, 8 September 2005 (UTC) ````

All of the documentation (e.g., end-user manuals, customer courseware manuals, and application progammer manuals) were written [by me, Walt V. Bernacki--Senior Documentation Specialist--Walt Bernacki now lives in Cape May, NJ; Assistant Professor of English at Atlantic Cape Community College] using a very primitive form of a word processor called "INTERACT," which was designed and developed by Cullinane Database Software programmers to simply provide a notation tool for the COBOL programmers to use as a means to capture design notes. INTERACT required the use of an IBM-3270 "dumb terminal" linked through dedicated data lines to Cullinet's mainframe located just north of Boston. Ads/Online was the first pliable software package that allowed database programmers to design end-user screens by actually using a visual "map" of each screen image that could be tailored visually to any database end users in a variety of technologies. Clients included Martin Marietta, GE Space, US Department of Defense, Missouri State Census and Payroll Records Department, and NORAD.

Question about Interact

Was not Interact a version of Wylbur?

In 1979, as a new employee, I was told it was a derived from the SLAC (Stanford Linear Accelerator Center) version of Wylbur.

Cullinane acquired the product (originally called "MENTEXT"), through the acquisition of Mentel Corporation.

Interact was used by Cullinane field personnel using TTY devices, not just corporate folk fortunate enough to have 3270s.

ADBADQC (talk) 21:34, 29 January 2009 (UTC)

ADS/OnLine Software was one of the first online application development tools.

All of the documentation (e.g., end-user manuals, customer courseware manuals, and application progammer manuals) for the ADS/OnLine product line were written [by me, Walt V. Bernacki--Senior Documentation Specialist--Walt Bernacki now lives in Cape May, NJ; Assistant Professor of English at Atlantic Cape Community College] using a very primitive form of a word processor called "INTERACT," which was designed and developed by Cullinane Database Software programmers to simply provide a notation tool for the COBOL programmers to use as a means to capture design notes. INTERACT required the use of an IBM-3270 "dumb terminal" linked through dedicated data lines to Cullinet's mainframe located just north of Boston. Ads/Online was the first pliable software package that allowed database programmers to design end-user screens by actually using a visual "map" of each screen image that could be tailored visually to any database end users in a variety of technologies. Clients included Martin Marietta, GE Space, US Department of Defense, Missouri State Census and Payroll Records Department, and NORAD.

ADS/O programs were easier to understand. It provided a "conversational" approach to applications development since it encaspulated the pre-processing, screen-design, and post-processing all in one module. (PreMap, Map (screen), Response.) Contrast this to CICS programs, and even today's Visusal Studio where the "conversations" crosses from one program to another. —Preceding unsigned comment added by TaxMax (talk • contribs) 07:50, 19 September 2010 (UTC)

Multiple Editor Personality Disorder for IDS origin

Article contains the following internal debate:

[...] taking over development of a Honeywell database management system called IDS that had been modified to operate on IBM and IBM compatible (RCA) mainframes . Actually IDS was originally developed by General Electric, and a Bill Curtis had supposedly gotten the rights to convert the system to run on IBM equipment.

Somebody needs to clear up its origin (with cite) rather than carrying on an argument within the article itself. Also, IDS is ambiguous (I've tagged it as such), and Bill Curtis currently links to a redir to Bill Kurtis, which I'm guessing is wrong... .

"a Bill Curtis" and "supposedly" are also quite unencyclopedic.--NapoliRoma (talk) 22:06, 22 December 2007 (UTC)

IDS was originally developed by Charles W. Bachman [[1]] when he was at General Electric for the Honeywell computer, which they owned at the time. He later ported the software from the Honeywell computer to an IBM mainframe for B. F. Goodrich, changing the name to the Integrated Data Management System, or IDMS). B. F. Goodrich subsequently sold all commercial rights to Cullinane Corporation in 1972. [[2]]Carlolofson (talk) 21:08, 13 November 2017 (UTC)carlolofson

and a disputed statement

Article also had:

The solution to the company's revenue problem turned out to be its new Integrated Data Dictionary [NOTE: THIS IS WRONG. IDD WAS PART OF IDMS WAY BACK IN THE 1970S.].

The "THIS IS WRONG" looks to have been added by Special:Contributions/99.233.218.21 on October 4, 2007. I've replaced it with a tag. --NapoliRoma (talk) 22:29, 22 December 2007 (UTC)

--

I agree. IDD was at the core of IDMS when I first used it in 1977 and there was no new "dictionary" in development as of 1987. —Preceding unsigned comment added by 24.141.39.66 (talk) 17:05, 3 August 2008 (UTC)

The "I" in IDMS stands for "Integrated", which refers to the IDD. I know IDD was part of the product when I first installed it in 1975. ADBADQC (talk) 21:39, 29 January 2009 (UTC)

MEPD III

Another debate, this time self-contained within one sentence:

In the process, Cullinet acquired some very questionable VAX companies but one had an outstanding relational DBMS but by then it was too late, the company's $50 million nest egg had been burned.

This one might be more a formatting problem than self-contradictory (although it's still quite loaded with POV).

While I'm at it, I might as well point out that "Goldengate was a mistake" is heavily into The POV Zone.--NapoliRoma (talk) 22:33, 22 December 2007 (UTC)

GoldenGate A Mistake?

GoldenGate was conceived at the time when Lotus 123 was dominant and Lotus Symphony just released. It looked like users wanted integration and Symphony was the first of many attempts to learn what integratiuon actually meant. Cullinet believed that its users wanted mainframe data to extend down to the desktop, and launched GoldenGate along with Information Center Managament System (ICMS). GoldenGate was launched before ICMS was ready and looked like a Symphony wanna-be. It was rather useless without the mainframe integration. When finally launched, ICMS proved to be far too slow to be of any use, and besides, the users nevery really figured out what to use the thing for.

Sinobra (talk) 15:15, 29 June 2010 (UTC)

I worked on the GoldenGate team. As I understood it, Cullinet's customers were making requests that ran something like this:

"We often want to generate reports from the corporate data held on the mainframe. So we send a request to the DP department for a report with the information we want organized in a table, with maybe a graph, and some explanatory text. If we are lucky, we can get our report in 6 weeks. Now that we have these IBM PC's in our offices, can't you supply us with a program that would allow us to download a subset of data from the mainframe, into a local database on the PC where we can sort, project, and select, and maybe transfer the data into a spreadsheet, and maybe generate a graph or two, and integrate it into a document with some explanatory text?"

GoldenGate was developed to meet this request. At the time it was classified as a "decision-support tool" as it was expected that reports generated by this process would be used by management in their decision making processes.

After the launch of the initial product, there seemed to be a lot of confusion as to what the next phase would be, and I think a lot of time and money was wasted.

I think upper management missed a opportunity here. Cullinet had great mainframe product and a enormously effective (and entrenched) direct sales infrastructure. Mainframe sales were the ultimate goal of the company, and Goldengate was created to enable more mainframe sales. It was the tail of the dog, so to speak.

I don't think Cullinet could conceive of the idea of spinning off GoldenGate technology as a separate product to be shrink-wrapped and mass marketed as a low-cost integrated office software solution. Had they created a division with this goal in mind, with it's own sales and marketing team, they might have been able to realize some serious income for some time. However, given Microsoft's determination to own the office automation software segment. It's doubtful Cullinet could have succeeded where Lotus failed.

End Sinobra (talk) 15:15, 29 June 2010 (UTC)

GoldenGate was not a mistake. Remember this was a DOS product. But it targetted only the existing IDMS Users. Not the oridnary popcorn PC user. And you needed a key to install. The IDMS mainframe link worked but LU2 based link had limitations of its own. Developments in technology obsoleted the product.

I agree: Goldengate was not a mistake. I was hired to help guide Goldengate into the Windows era. We were inspired by Apple and HP with their iconic interfaces and interations. But by that time it was really too little too late. — Preceding unsigned comment added by 198.23.5.10 (talk) 16:32, 20 September 2011 (UTC)

Slang

What does "the company had got hung up on the open architecture and relational issues" mean?Lestrade (talk) 00:33, 24 June 2008 (UTC)Lestrade

I guess it means there were too many technological changes at the time for a product suite entirely written in S/370 & S/4300 (that is, strictly Mainframe) Assembler Language to catch up with. Going to relational was an immense effort, but trying to cope with newly emerging platforms literally caught Cullinet with their pants down. --AVM (talk) 18:02, 7 April 2009 (UTC)

IMHO, this means that IDMS cold never be positioned as a relational database. IDMS was a network database (some would say Codasysl). Being network oriented meant that one could only construct relationships using pointers within linked lists: Next/Prior/Owner). This is certainly not the same as the more unstructired JOINs available in a realtional database.

IDMS ran much faster than the original release of DB2. It should have been easy for marketing to neutralize the erosion DB2 made against IDMS's market share. IBM, on the other hand, did an excellent job of making relational a must-have for many corporations. —Preceding unsigned comment added by TaxMax (talk • contribs) 07:42, 19 September 2010 (UTC)

Well, IDMS/R was the mistake. This was not a relation DBMS as such but just an LRF (logical record facility front end) layer on top of teh CODASYL database. "R" in IDMS/R confused the market and even technicians and did all the damage alongwith the infamous comparison of 12 rules of EF Codd. IDMS did come with a proper relational database in 1992. But it was too late when the mainframe itself was under threat. I do not believe DB2 as such was the reason for IDMS's market failure in the longterm. The issue was platform. Later DB2 itself lost the market to Oracle. And today in 2011 the industry is discussing E-R Domain Models, ORM (Object relational Mapping) and NOSQL Databases!! IDMS which is a 1:1 implementation of E-R Model is easily better suited to represent today's Java Objects! One doesn't need any ORM here!

Open Architecture?

That was never an issue.

Cullinet got hung up on relational. IBM began agressively marketing DB2 to its mainframe users. The positioning was that relational database systems were the future and anything else was obsolete. They even suggested that companies cuold replace their programmers with DBS tolols such as QBE. It was pure hype.

Cullinet, in a vain attempt to preserve sales, tied to position IDMS as a relational-like database sytem without defining what that meant. E.F. Codd was right to berate Cullinet for this. Im my opnion, Cullinet should have positioned IDMS as the database for large enterprise applications because it easily outperformed DB2 by an order of magnitude. I think that the fact that there remain many large-scale IDMS databases are still in use in 2008 proves this. —Preceding unsigned comment added by 24.141.39.66 (talk) 17:36, 3 August 2008 (UTC)

Too bad you weren't there at the time, to assist Mr. Cullinane make better decisions. --AVM (talk) 18:09, 7 April 2009 (UTC)

I was there at the time -- but being Canadian, my input was not valued. —Preceding unsigned comment added by 24.150.160.127 (talk) 01:56, 2 April 2010 (UTC)

Well, again Cullinet gave up too early! Cullinet Management was fooled by the market perception regarding the "death of mainframe" which did not happen! And 10 years after being sold to CA, the majority of IDMS sites were still in production in 1999 and working on Y2K! Products like Oracle were later helped by Java which was not even born when Cullinet was sold! Even today IDMS database has the ability to implement both CODASYL and Relational Models. When the industry is discussing NOSQL databases and storing of unstructured data (JPG, PDF etc), IDMS was already doing it in storing the run time load modules (binary) in dictionary load area. Vegasoft product from Helsinki provided Web Access to IDMS and was able to store all the artifacts of "html" (images and html and xml etc) in the IDD!

ADS/Batch

Initially, ADS/Batch was a reworking of a package purchased from a customer and did not have the elegance of ADS/Online. The developer in charge of reworking the product left the company shortly after the product was released because she knew that she would forever remain infamously associated with the product. Only later did ADS/Batch get rewritten as an ADS/Online port. TLayman 22:55, 26 February 2010 (UTC) —Preceding unsigned comment added by Tglayman (talk • contribs)

She was smart. Very few people inside Cullinet even understood the product. —Preceding unsigned comment added by TaxMax (talk • contribs) 07:44, 19 September 2010 (UTC)

ADS/Batch was bound to be a failure. The plus point of ADS/Online was the structure to make easy ONLINE applications where transactions are short lived and accesses are of sub seconds. ADS code was not fast as COBOL. But this did not matter in an OLTP situation where the response time = elapsed time + code CPU + DBIO. Here code CPU is negligible compared to the rest. But in a batch, the CPU time plays a major role because the batch is run on 100s of 1000s records or even more! ADS Code written in 1984 is still in active production in many parts of the world!!

Cullinane left PHI to form his company

Perhaps it should be noted that prior to forming Cullinane/Cullinet that John Cullinane spent a number of years working for PHI in Arlington Mass before going out on his own after the aquisition of PHI by Wang Labs.

PHI was the premier software development house in the region, clients ranged from Draper Labs, IBM, to commercial accounts such as Marine Midland bank —Preceding unsigned comment added by Jv818 (talk • contribs) 20:11, 29 November 2010 (UTC)

History lead desert

"History" is such a "lead desert", could anybody please reformat it and divide it in smaller sub-paragraphs? As it is NPOV-disputed, I sat the "needs-attention"-flag.

Too subjective

This article is littered with unsubstantiated subjective commentary. It should be cleaned up.Carlolofson (talk) 21:12, 13 November 2017 (UTC)carlolofson