schultz@grebyn.com (Ronald Schultz) (02/09/90)
Many consultants are stating to us poor MIS types that object- oriented analysis and possibly design are implementable in even the unholiest of languages, ie. COBOL. Is an object-oriented design implementable in COBOL? If so, how? The solutions I see are either extremely inefficient to the point of unworkable, or so complex as to not be worthwhile. Ny thoughts would be sincerely appreciated. No flames please and serious replies only. I don't need any language bigots going berserk right now! I'm sorry about discussing COBOL in COMP.OBJECT, but 70 billion lines of code can not be dismissed by a paradigm shift, and if OO is so great, somehow there has to be a means of demonstrating this superiority to an MIS shop, even if it is in COBOL. Ron Schultz schultz@grebyn.com
megana@hpcll19.HP.COM (Megan Adams) (02/21/90)
There is a CODASYL Taskgroup investigating Object-Oriented COBOL. (CODASYL is the committee which has the charter to develop new features for ANSI Standard COBOL.) I'll respond via email to anyone wishing information on this Taskgroup. Membership is open but entails a work requirement; observers at Taskgroup meetings are permitted.
davidc@vlsisj.VLSI.COM (David Chapman) (02/21/90)
In article <1990Feb15.004231.23324@alzabo.uucp> tris@alzabo.uucp (Tris Orendorff) writes: | One of the benefits of OO is being able to share code. MIS |shops don't know the meaning of the word share. That's why there is 70 |billion lines of code. Everybody started out from scratch 20 years ago |and has build up proprietary routines to get their work done. | This may also explain why there is no comp.lang.cobol or |alt.cobol for people to share COBOL information. The computing centers |are all closed environments and will probably stay that way until they |are shut down (fat chance). There is an alt.cobol, by the way. I read the first few messages because I thought it was a joke, but it's real. I don't know if there's a comp.lang.cobol, though. We don't have either here, but Stanford has alt.cobol. Be still my beating heart. :-) -- David Chapman {known world}!decwrl!vlsisj!fndry!davidc vlsisj!fndry!davidc@decwrl.dec.com
gph@hpcc01.HP.COM (Paul Houtz) (03/01/90)
tris@alzabo.uucp (Tris Orendorff) >schultz@grebyn.com (Ronald Schultz) writes: > > >>Many consultants are stating to us poor MIS types that object- >>oriented analysis and possibly design are implementable in even >>the unholiest of languages, ie. COBOL. Is an object-oriented >>design implementable in COBOL? If so, how? >> ... > > I think that it would be possible to add OO to COBOL. However, >my real question is: Will anyone use it? > One of the benefits of OO is being able to share code. MIS >shops don't know the meaning of the word share. That's why there is 70 >billion lines of code. Everybody started out from scratch 20 years ago >and has build up proprietary routines to get their work done. > This may also explain why there is no comp.lang.cobol or >alt.cobol for people to share COBOL information. The computing centers >are all closed environments and will probably stay that way until they >are shut down (fat chance). > > Sincerely Yours > Tris Orendorff > tris@alzabo.uucp Tris, you have worked in some interesting COBOL shops, I guess? I have worked in and with a fair number of COBOL based MIS departments. My experience is that COBOL programmers are actually just the same type of people as any other programmer (a revolutionary idea, I know), and that many of them understand the value of code sharing. In fact, I was more likely to steal COBOL code than to write it because it is so typing intensive. There is an incredibly obvious reason why there are few COBOL notes groups; Very few COBOL environments are Unix environments, so they are not going to be connected to the network. It is simply not true that COBOL computing centers are closed environments; they simply are not connected to the particular network you rely on. Actually, many of them are converting to unix and using such products as the MicroFocus COBOL Compiler to do development and production on Unix systems. MF COBOL is used by SUN, HP, and other vendors as their COBOL solution for the business computing environment. I expect you will see more COBOL, not less, in the future. As for the object oriented use of COBOL, I myself can see no reason why object oriented priciples used in COBOL would have poor performance. I would think that COBOL would be an IDEAL language for Object Oriented design. It is difficult to pass many different data types to sub programs in COBOL, and it is much easier to pass messages in the form of strings. Also, the Paragraph Structure of COBOL is a very low overhead implementation that would allow a subprogram to have many private routines to use to access a data structure, and hiding of the data structure is easier in a COBOL subprogram than revealing it.
tris@alzabo.uucp (Tris Orendorff) (03/05/90)
gph@hpcc01.HP.COM (Paul Houtz) writes: > I have worked in and with a fair number of COBOL based MIS departments. > It is simply not true that COBOL computing centers are closed environments; > they simply are not connected to the particular network you rely on. This is possible. > As for the object oriented use of COBOL, I myself can see no reason why > object oriented priciples used in COBOL would have poor performance. > I would think that COBOL would be an IDEAL language for Object Oriented > design. It is difficult to pass many different data types to sub > programs in COBOL, and it is much easier to pass messages in the form > of strings. I agree. On the question of performance, I don't think it would be poor either. OO principles have more benefits than liabilities. I can't wait to see what the new COBOL will be like. I suspect it may be similar to some of the fourth generation languages that are already out. -- Sincerely Yours Tris Orendorff tris@alzabo.uucp ----------------------------------------------------------------------- ALGERNON: The doctors found out that Bunbury could not live ... so Bunbury died. LADY BRACKNELL: He seems to have had great confidence in the opinion of his physicians.
wdr@wang.com (William Ricker) (03/06/90)
For those interested in this topic, Ed Yourdon's "American Programmer" magazine's March'90 issue has "OO COBOL" as its cover article. Check a corporate library near you, or attend a conference Ed is at; he seems to give out lots of free sample copies to stimulate subscription. (No, you can't afford it personally: its UA$295/yr.) The article explains why this apparent double-heresy shouldn't be surprising. -- bill ricker wdr@wang.com