On migration

This is the point. From that person's perspective they were faced with high risk (to them) situation. They see change in programming language from what they feel comfortable with using and/or managing.

And from the Geoffrey Moore explanation they were part of the "Early Majority" who always ask their peers what they are using. And it would not be Lisp.

You yourself face this with Cincom Smalltalk.

Comment by James Robertson

Alex,
I still maintain that porting from working, maintainable software to something else is a waste of time and resources.

Depends on your definition of "maintainable".

Had you inherited Bottom Feeder already written in Perl, or VBScript or (well you get the idea) what would you have done?

This person, I suspect, saw Lisp as "not maintainable" in his management view (leaning on the views of his peers).

Comment by James Robertson

Alex - he had Lisp developers in house, unless he fired them. And - regardless of what you seem to think, it's not that hard to find new ones or train existing staff. Certainly it's far, far less of a hit than rewriting the entire app.

In base business terms, the decision was unjustifiable. Move it out of software, and it's obvious how stupid it was.

James: "because some thundering moron decided"
Do you have any inside information on 'who decided what' and why?
Or is this name-calling based-on ignorance?

I am not arguing that it was a good business decision. It is not what either of us would have done. I only say that this is the way the Early Majority will act. It is quite understandable behavior given the Early Majority "risk assessment and avoidance" strategy.

From understanding comes a plan (of how to grow Smalltalk (or Lisp))

Comment by James Robertson

Isaac - IMNSHO, if you have working software, and all it would take to keep it going and enhance it is training/hiring some staff who know (Insert niche language here) - as opposed to doing what these people did - rewriting from scratch - you are are ipso facto dealing with a moron. To be blunt, there is simply no business level justification for that. It's like finding that your office was built out of wood, declaring that unacceptable - and razing and rebuilding with steel.

Comment by Vincent Foley

Had you inherited Bottom Feeder already written in Perl, or VBScript or (well you get the idea) what would you have done? This person, I suspect, saw Lisp as "not maintainable" in his management view (leaning on the views of his peers).

There's a difference. If you move from VBScript to Smalltalk, you move up the "flexibility" ladder, so your application can have features otherwise not possible with VBScript.

The problem with Yahoo! Stores is that the Common Lisp version had better, more advanced features than what they got in C (for example, in Lisp, pages could return a value, like a function.) This means that the overall user experience is decreased to please management. That's the big deal. I am not against rewrites if those bring advantages to the users, if they have a better system. I'm not for doing it for political reasons though.

Vincent, yes and no!

No argument that Lisp and Smalltalk are clearly superior languages, that is amongst us "Early Adopters". The point I am trying to make is please stop and understand the mindset of the Early Majority. They attribute different values to the parameters of the risk-reward equation, thereby deriving a different solution. You will not change that, they do not believe they are wrong, so I am saying understand and deal with it!.

OTOH, a well written VBScript should not defacto be replaced with Smalltalk, or whatever your favorite language is. You must examine the return on investment. And of course this equation is neither simple, nor could we all agree on the values of the parameters.

Now I must say that I would have a tendancy (though it is not a given) to rewrite in Lisp or Smalltalk (or a few other interesting language/development environment solutions). But then I am an Early Adopter, and my equation parameters are different.

James: "To be blunt, there is simply no business level justification for that."
To be blunt, you haven't a clue if there were business level justifications or not.

Comment by James Robertson

Isaac - the simple fact is, it would be far cheaper to either train a new developer in Lisp, or hire an experienced one than it was to rewrite. Go ahead, find me a business justification.

James,
It appears you discard the "perceived" risk -- see above.

Comment by James Robertson

Alex, I understand the perceived risk - but that perceived risk never had cost/benefits analysis applied to it

James,
I am extrapolating here, but it is my guess that you do not (truly) buy the Geoffrey Moore concept?

The behavior of the "Early Majority" is not irrational, moronic, stupid... It is the way most people behave. (The outcome may be stupid, but then we can point to lots of that in all aspects of life.) They just have different values for the parameters for their equation of risk/reward, different experiences and knowledge and different peers to influence them. And you are probably correct that they did not break out Excel. Most of us make gut/intuitive decisions based, perhaps, on years of experience, and/or discussion with peers. There is just not enough time for us to break out a spreadsheet for every decision. I will bet even you make some decisions intuitively, based on experience, where you should have done some cost/benefit analysis?

Just imagine that everyone you respect tells you "don't do that". It is very difficult to ignore, and most people will not ignore! We like to belong. It is the rare individual that does not "follow the herd".

I am not saying this is good. Just that it is so. And understanding it helps us deal with it. My belief is that if Cincom internalized this they could use it to grow Smalltalk to where it should be in the world (which is the only reason I spend time here - selfishly).

Comment by James Robertson

Alex,
I understand Moore's points very well - and I understand the underlying behavior. All I'm pointing out is that in this kind of case - where you have a functioning application - rewriting it in order to be more comfortable makes no sense. It's expensive and wastes corporate resources...

I should probably quit - this is the last on the subject.

Would you refactor a functioning application? Would you put investment into an application that is already functioning to make it in your judgement more maintainable?

You can have the last word.

James: "Go ahead, find me a business justification."

The Principal Forms of Illogical Thinking
No. 22 "An Inability to Disprove Does Not Prove"

Being Logical: A Guide to Good Thinking

Comment by James Robertson

Isaac,
Bzzzt - nice try, but that's not a business justification.

James, you win!

You bad-mouth someone you don't know on the basis of gossip; and you think the onus is on others to disprove your idle speculations.
Not worth my time.