Review of D-Day 2000
Shawn M. Gordon
It’s on the front page of all the trade rags, “the year 2000 is coming fast, are you ready”? I’ve written a bit on this topic in the past, and even reviewed one package so far, with at least two more to come after this one. Basically back in the 60’s, 70’s and 80’s when most code on the planet was written, no one ever envisioned it running for more than 5 to 10 years. Since disk was expensive it was important to conserve ever possible bit of space, so everyone stored their dates with no century, after all, who cared, all the code would be gone soon anyway.
Well they were wrong, and the year 2000 is probably the most lucrative business on the planet right now. The problem is that changing all this code requires extensive research and manual work, so this is why we are seeing such a sudden dearth of tools. This is where D-Day 2000 comes in (I love that name). This product is an extension of Diamond Optimums fine DOC/3000 tool which will do it’s best to document your systems for you.
I tested D-Day 2000 on a bunch of COBOL applications that I knew had date problems on my 925.
The first thing you must do when setting up D-Day 2000 is get it configured for Doc/3000 to get indexed. Take a look at Figure 1 for the basic project setup screen. This is where you can define what DOS calls a System, and relate your files to a particular system, for me this is very handy because I have a dozen different products that all reside in the same area, and it’s nice to be able to work on them as distinct entities.
It’s important to note here that you can check anything you want, not just source code, and not just COBOL. That include COPYLIB’s, JCL, INCLUDE files, etc. This is a full featured system for documenting.
Once you have your system set up and configured, you launch the standard DOC/3000 collection job, and let it do it’s duty. After DOC/3000 has collected all of it’s information, you will then want to use the D-Day 2000 section to configure, and run the types of reports that you are interested in. Basically you have and INCLUDE filter file, and an EXCLUDE filter file that you can manipulate so you don’t end up with too much data on your reports. I found that you will probably need to make a couple of run through’s before you find the balance you need. The first time through I had WAY to much code qualify based on the Include filter default setup.
There is another option to convert the DDAYLIST to a DDAYFILE file, but I couldn’t find any documentation for it. From what I could tell from the Client software, you can use this file to do date fixes on code, but I couldn’t verify it.
D-Day 2000 is basically a batch tool where you generate a bunch of reports on projects and then go and fix your code. You can use the client software to give you sort of a bird’s eye view of the (impact analysis) on your systems, and what is going to be involved, see figure 2 for an example of the graph of complexity matrix by application. You can see that FYIMAIL is much easier than JOBTRAK. Figure 3 show’s us the Impact statistics by application matrix, the color chart at the top tell’s us what each color on the graph is for. This drops down when you click the button, it isn’t always hanging over your display.
The reports include “Application Level Impact Analysis and Metrics”, this is similar to the graphs in that it describes the total impact by system, and show’s no detail. The “File Level Impact Analysis and Metrics” report shows each program and statistics on it as well as the item’s that matched the search criteria, with the number of occurances. The “File Level Metrics and Complexity Matrix” report shows each impacted file, the System it is part of, the language, and line number, the year 2000 impact side of the report shows the type of code that was involved, MOVE, CALC, COMPARE, etc. Finally the Line Level Impact Analysis report shows the actually code in each affected program as well as the word that matched in the code. This last report is the most useful for getting to the nitty gritty.
The client software for the Doc/3000 portion is pretty fun to use, and useful on top of that. Take a look at figure 4, this allows you look for certain “entity types” within different files, and then be able to pull up the file and view it as well as search for strings within it.
The last thing I am going to go over is the client software impact analysis, see figure 5. This allows to search by elements within source code and find what qualifies for a particular object search type. This is handy for jumping around, and relieves the necessity of possibly perusing through what could be a rather large document.
It’s important to note that you end up with a copy of DOC/3000 as part of getting D-Day 2000, this add’s quit a bit of value, and long term usefullness (beyond 2000) to the package.
Usability (also installation)
The installation from Diamond Optimum is a mature and solid process, and goes very easily. The software is quit easy to navigate through, but sort of suffers from a dual personality due to it’s mixed use of block mode and character mode screens. This makes it slightly less intuitive and requires that you check out the prompts before responding. Once your set up though, which could take a few tries, things run very smoothly and easily from there on.
D-Day 2000 run’s very dependably, all the online processes execute without fail, and the reports are complete as far as I can tell. Usually as a litmus test of a products idiot-proofness, I will muck around for a while without looking at the documentation to see if I can mess up. I did this with D-Day 2000, and was not able to cause a failure. High marks for solid software.
The reports are very very fast in executing, I was able to generate reams of data quit quickly. I say reams, because my first time through showed me that I needed to set up more “exclude” keywords.
The online process is trivial in it’s impact, you are just entering a few parameters in screens.
Supportability (including Doc)
Support from DOS is alway of good quality, and I have always been happy with it. My biggest complaint with the doc is lack of a tutorial guide. There is a bit of a walk through, but not really a full fledged tutorial, I really think that this type of product needs to have a small sample that will walk you through the whole process of setting up, running reports, fine tuning the system, running more reports. It’s a little tedious, but it allows you to take full advantage of the software and not muddle around.
As always, the complaints come first. Basically the online portion of the product has the feel of something that was knocked out quickly. The whole manipulation of the include and exclude files should be handled in the online software, and shouldn’t require you to use editor and fcopy on your own to maintain them. The reports on the other hand have the feel of something that was pretty carefully thought out and coded, they are very informative, and well done.
You could do part of what D-Day does by using various utility programs out there, but you would be going all over the place, and wouldn’t get the informative reports. I think D-Day is a great tool for identifying problems, then something like MPEX for making mass changes to code, and then something like Adager or DBGeneral for manipulating the data and databases into the new format. This is what you need to get through the problem as quickly as possible, with minimal expense. I would recommend that you check out D-Day soon, or it really will be D-Day
D-Day 2000 version
Diamond Optimum Systems
22801 Ventura Blvd. Suite 105
Woodland Hills, CA 91364