Why Government Computers Are Such a Mess: Hercules Has Nothing On Musk’s Wizards

Every so often, I get reminded that I’m old, and I’ve been programming for about 60 years, which is a long period. However, having spent 60 times in the industry, I’ve seen many things that even the most naive developers have never seen.

This is frequently brought up when DOGE and the Wizards Academy Musk has created to aid in the investigation of fraud, misuse, and, most importantly, administrative foolishness are discussed.

One of the things that fresh, technically educated people say about things like Social Security and the IRS is that they should simply dump the entire collection into Hadoop.

The issue with that begins with the fact that it isn’t stored in a collection. It’s a madly varied collection of various databases, ISAM data, and card pictures, and I would imagine money that a lot of it is on older 7-track tapes. Some of these are most likely housed in Iron Mountain or another construction. Moreover, some of the data may still be only on paper, as, obviously, state pensions records are.

But what Big Balls and another wizards will need to do first is gather the data. &nbsp,

I’m willing to bet there’s no one archive of all the data set. After discovering the information, the majority of it is card images, which are almost undoubtedly just COBOL copybooks’s ability to document. &nbsp, ( Back in 2020, &nbsp, I wrote about COBOL&nbsp, for the Stack Overflow Blog when it suddenly became trendy as some of these systems desperately needed to be maintained. )

A COBOL functions is essentially a copy of the code that is physically stored in a different location and “included” in additional programming languages, similar to what is known as an “include file.” In COBOL, a functions might look like:

01 RECORD. &nbsp,
&nbsp, &nbsp, 05 ID &nbsp, &nbsp, &nbsp, &nbsp, &nbsp, &nbsp, &nbsp, &nbsp, &nbsp, PIC S9 ( 4 ) COMP. &nbsp,
&nbsp, &nbsp, 05 COMPANY. &nbsp,
&nbsp, &nbsp, &nbsp, 10 SHORT-NAME &nbsp, &nbsp, &nbsp, PIC X ( 10 ). &nbsp,
&nbsp, &nbsp, &nbsp, 10 COMPANY-ID-NUM &nbsp, PIC 9 ( 5 ) COMP-3. &nbsp,
&nbsp, &nbsp, &nbsp, 10 COMPANY-ID-STR&nbsp,
&nbsp, &nbsp, 05 METADATA. &nbsp,
&nbsp, &nbsp, &nbsp, 10 CLIENTID &nbsp, &nbsp, &nbsp, &nbsp, PIC X ( 15 ). &nbsp,
&nbsp, &nbsp, &nbsp, 10 REGISTRATION-NUM PIC X ( 10 ). &nbsp,
&nbsp, &nbsp, &nbsp, 10 NUMBER-OF-ACCTS &nbsp, PIC 9 ( 03 ) COMP-3.

This is derived from a GitHub store and is licensed as open-source by Apache.

There’s an extra problem: most computer information is encoded in a different coding standard called EBCDIC than what’s used frequently in the rest of the world, known as ASCII.

There’s no reason why you can’t use a COBOL functions to create password in some more hacker-friendly language, which might be a job that might be partially passed on to some AI, but you still need to evaluate the script, which will take some time on its own.

The point is that not even Major Balloons will be able to complete it in a few days.

The reason DOGE needs to do this is to conduct an exact assessment, or one of the causes. We already know that this information is a disaster, as we frequently say in business. ( Actually, we would be more vehement than that, but Google doesn’t like it when we swear in an article. )

Actually, it occurs to me that some of the information may not be described in a copybook, but only in the COBOL or PL/ I or HLASM source code. Therefore, it would be necessary for someone to read and record down the script.

Anyway, as I say, we already know that the information is a disaster. What is known as a” Social Security account number” ( SSAN ) is what we call it. It does, by definition, determine one living person truly. The Social Security Administration claims it never uses an SSAN, but in 2011 it began arbitrarily assigning new figures within time limitations because they were running away.

The general idea behind getting a Social Security number is that it has the well-known script 123-45-6789 with a few extra rules, and it should be entirely theirs. This is an interesting story merely for nerds, especially security nerds.

Still, there are Social Security numbers that can be assigned to a total of eleven people.

How can that happen? Easy: they f***ing fouled up. ( Darned Google. )

When this was first reported, a lot of people said,” That can’t happen”!

Except, presumably, it does.

Now, if the data were kept in a single database, records for a new user with an SSAN would soon display an error message, and the new document would be discarded to be physically fixed.

Obviously, they aren’t, and at least part of the reason is that we don’t really have a consolidated collection. &nbsp,

To start with, what we now think of as a “proper” relational collection was simply invented in 1970 and didn’t get into popular usage until the late ‘ 70s, with IBM SQL/DS in 1976 and Oracle Version 2 in 1979.

Social Security numbers, but, started being issued in November 1936, and the SSA just started computerizing in 1956, with big digitalization starting in 1961.

According to Musk, there are people who were reportedly born around 1875 and who were on Social Security floats as ancient as 150.

Now, one might be cautious that this was really done on purpose. I’m willing to bet it wasn’t. However, it was done by people who didn’t have a full picture of the entire system, and I bet there isn’t anyone who does, making changes to control anything unusual that occurred and getting it done by Tuesday because their quarterly assessment is upcoming.

One fact in computer safety, though, is if there’s a weakness that can be exploited, it will be exploited. It’s pretty obvious from what the Wizards Academy has already discovered that everyone has been paying it, including the man who keeps cashing great-granddad’s Social Security investigations because, oh, the government sent them to effective fraud rings.

Hercules did consider cleaning the Augean horses was a breeze if they cleaned it up.

It may actually get worse.

Source credit