Intel Corporation
350 East Plumeria; Mail Stop CHP3-105
San Jose, CA 95124
408 765 8080


Date: Tue, 8 Jan 2002 17:55:44 -0800



Mr. Garold L. Johnson
dynalt@dynalt.com
Dynamic Alternatives

Subject:   SDS upgrade

Gary,

[Responding to your letter earlier today... ]

Please see my answers in Blue...

[Gary Johnson wrote....]

Morris,

Rod has talked with you about the possibility that I might help with completing the task of converting MEDIT to C using the Windows model to increase file sizes, etc.

Before I commit to this, however, I wanted to discuss several of the issues that I see involved.

Why don't you do it yourself?

[Jones, Morris] I haven't found/made the time. SDS is both a management method, and a software platform. I am also employed and trying to raise kids, stay married, have a life, etc.

[Gary Johnson wrote....]

Rod speaks glowingly of your capabilities. You wrote the program in the first place and started on the C version.

[Jones, Morris] I wrote an editor with macro capabilities. Rod turned it into SDS through years of hard work. I used the C version personally as an editor. The editor is like the old IBM VM editor.

[Gary Johnson wrote....]

A reasonable question is, therefore, why aren't you doing the rewrite yourself? You are clearly far better qualified than any newcomer can be.

[Jones, Morris] See above.

[Gary Johnson wrote....]

Also, a major part of the attraction to doing this at all is the possibility of collaborating with someone from whom I stand to learn a great deal. If you don't have time to work the problem, it is likely that you will also be short on time to collaborate on the work, thus losing a major attraction for me.

You don't use SDS yourself?

[Jones, Morris] No

[Gary Johnson wrote....]

According to Rod you started to use SDS at some point and gave it up. May I ask why?

[Jones, Morris] It was too hard to use in those days. My work style didn't lend it's self to long sessions of analysis.

[Gary Johnson wrote....]

At the moment, it seems that Rod may be the only person who can use SDS effectively.

[Jones, Morris] There are a few others, but the list is small at this point.

[Gary Johnson wrote....]

Project goals (as I see them)

Rod believes that the objective is to rewrite the existing MEDIT, including the macro language, in C, and thus have SDS move with the conversion.

[Jones, Morris] He doesn't want to rewrite his macros. I have suggested he learn C and then port the macros to that, but he hasn't done it to date.

[Gary Johnson wrote....]

First, is a direct rewrite adequate? Any time I have set about to rewrite a program, there have been numerous things that I wanted new or to do differently. I assume that MEDIT is no different.

[Jones, Morris] I feel in the end you will really want a full blow application. A text mode MEDIT on windows with more memory will solve some of Rods short term issues, but will not result in a very marketable product in my opinion.

[Gary Johnson wrote....]

Is MEDIT of interest beyond the fact that it supports SDS? Do you have plans for it that would levy additional requirements on the next version?

[Jones, Morris] No

[Gary Johnson wrote....]

Even assuming that an editor is the place to start in reconstructing SDS, is MEDIT the best starting point? There are several commercial editors that support a wide variety of scripting languages, compiled languages or even DLL's to enhance the editor. Wouldn't this save a substantial amount of energy in development?

[Jones, Morris] No, it's just the one Rods macros are currently developed in. I have suggested MS Word could be used with visual basic, but he hasn't gone there. Originally, he focused on speed as an important part of ease of use. Medit is written in assembly language, uses linked lists, and is quite fast on slow machines. In the days of 2GHz machines, that limit could be relaxed quite a bit.

[Gary Johnson wrote....]

There is also the possibility of treating the current version of SDS as a "proof of concept", as the "prototype to throw away". Perhaps it would be better to start with the current program as requirements and redevelop SDS from scratch.

[Jones, Morris] That is a reasonable starting point. I feel a good data flow and structure map needs to be created. And a different application created from those.

The current program could be used for that. I have tried this with Rod, but we didn't get very far.

The current program has a lot of dependencies on line numbers. This may require making some fundamental changes in the user interface and linking system to get rid of them.

Rod likes "grocery store codes", and uses them quite effectively. His subject index is much like a work breakdown system with a long string of small numbers representing an entry in the system. I feel you should explore how to accomplish the same things differently, and in a manner more easily adopted by a larger audience.

[Gary Johnson wrote....]

My status

I do not have much in the way of free time, and it is getting worse. My current contract will end within a few months at most, and I need to try again to establish a source of income.. Any effort I undertake on SDS will be done rather slowly.

[Jones, Morris] I understand. At this time, I am not aware of any formal funding for a SDS project.

[Gary Johnson wrote....]

I am not looking forward to going back to C, even with Windows. I would really rather look more toward the future than the past.

[Jones, Morris] What do you consider the future? I thought the language battles had died down somewhat. If developing for cross platforms, there is only C and Java. Java looks like C to me (Joke)

[Gary Johnson wrote....]

Do you have any insights into either SDS or the rewrite problem that would shed more light on the subject?

[Jones, Morris] I have quite a few on the editor core. I also understand the macro language Rod has used. He has the knowledge on the "SDS" use of the system. It will probably take both of us to really close the loop.

Sincerely,

Cable Network Operation



Morris E. Jones
Business Unit Manager
morris.jones@intel.com
408 765 8080


Copy to:

  1. Rod Welch, rowelch@attglobal.net