morethanabitoff

Java will be the New COBOL

Wed, 25 May 2016 05:20:06 -0400

We can continue to debate this, but Java will be the new COBOL.

The “Java is the new COBOL” debate has been going on for about 10 years now. And it’s not just theoretical. There are companies catering to what they sense is a shifting mindset in the enterprise programming world – a mindset that starts with a premise many of us would probably find unfamiliar: its time to switch from COBOL to Java.

If you want to know what I’m talking about, do a Google search for “Java for COBOL programmers”. You’ll find lot’s of resources for the COBOL programmer to learn Java. That highlights two things: 1) there are still quite a few COBOL programmers out there (stunning, right?), and 2) there is interest in moving them to Java, a language many of the cool kids now find passé.

The Scope of COBOL Today

COBOL is an old, old programming language, surfacing sometime around the cooling of the Earth’s crust, as the legend has been passed down from generation to generation around the evening campfire. So to say that there are still COBOL programmers can be surprizing.

So how many are there? Umm…

An estimated 2 million people in the world are currently working in COBOL; rough estimates put the US figure at around 20,000 people

Of course, there’s no source for that, but the other stats seem to indicate there’s a lot, A LOT, of COBOL still running around out there:

Some 75% of the world’s businesses data is still processed in Cobol, and about 90% of all financial transactions are in Cobol, according to Arunn Ramadoss, head of the academic connections program at Micro Focus International PLC, which provides software to help modernize Cobol applications.

Surprizing? Given that there is so much COBOL out there and this aging language’s aging practitioners aren’t getting any younger, some might say ALARMING. Ritika Trikha is predicting an impending COBOL gap:

But what’s not emphasized often enough is there’s a slowly increasing gap between the number of massive institutions relying on COBOL and its relevancy among programmers today.

And that maybe true if the law of supply and demand can be applied to COBOL programs needing COBOL programmers, as ITWorld has proclaimed: “College students who learn COBOL make more money”.

Out: COBOL. In: Java

One might conclude, given everything I noted above, that COBOL is the new COBOL… or maybe that Java cannot be the new COBOL because COBOL is still king. But looking at the trends, Java may not be the new COBOL now, but it will be one day.

Here’s one way to think about it. COBOL is what it is today because of the millions of man-hours that went into creating what we now call legacy business applications. Just think what hundreds-of-millions of man-hours of Java coding on today’s business applications will yield 30 years from now?

I say hundreds-of-millions because for the last decade or so Java has been (and continues to be) the most popular language in the world:

Java’s dominance on the TIOBE Index goes back to 2006.
It currently ranks #1 on the PYPL Index.
On both of the Javascript-biased* rankings, RedMonk and GitHut, it ranks #2.
It ranks #1 on IEEE Spectrum’s rankings.

There are probably an order of magnitude more Java programmers now than there were COBOL programmers back in the day. And therefore the man-hours being put into code are far greater.

But numbers don’t tell the whole story. COBOL’s dominance didn’t come from being the best programming language of the day but from targeting the average programmer. Rockstar programmers never fell for COBOL, but the vast majority of coders are not rockstars; they’re more like Karaoke singers.

The same is true of Java today. It isn’t elegant like Ruby, and it doesn’t have the same level of functional ease like Elixir. But your average programmer tends to make a mess out of dynamic languages and is certainly not prepared to handle functional paradigms.

So where COBOL was geared toward report generation of fixed record manipulation and report generation, the big tasks of the mainframe era, Java (and importantly, its many frameworks) is geared toward the middle-tier problem of manipulating relational databases on behalf of a GUI (be it web, mobile or a REST endpoint).

As C2.COM puts it:

Is Java the new Cobol? A good argument can be made that yes, it is. Considering the billions of lines of Cobol code that have been (or still are) in production; that observation should in no way be considered a slur. Java has succeeded in the marketplace (partially) because it is a language of its time; not one behind the times - or ahead of them.

*Notes and Asides

Javascript-biased Indexes) I say that RedMonk and GitHut language rankings are biased towards Javascript because they are based on GitHub code repository measurements, making them biased for two reasons.

First, as RedMonk points out, Javascript is used in the HTML of many of other projects where another language is really the dominant focus. So this inflates Javascript project and lines-of-code counts.

Second, the popularity of GitHub Pages likely inflates the Javascript numbers. Since all GitHub Pages are all GitHub repositories, any GitHub Pages site that might be about some other subject (take this blog for instance) would include Javascript files and thus inflate the counts.

The Mind Boggles with the Confluence of Javascript and COBOL) While we are on the topics of Javascript and COBOL, this article at Ars Technica on COBOL for Node.JS and vice-versa is interesting.

A cooling IPv4 Transfer Market

Tue, 24 May 2016 08:20:06 -0400

RIPE Labs on their view of the IPv4 transfer market:

In the second half of 2015 and the first half of 2016, both the number of transfers and the size of the transfers conducted under RIPE Policy started to show a downward trend. This is more likely a sign of fewer available addresses rather than decreasing demand for IPv4 addresses. Also, we have to stress that statistics for IPv4 transfers within the RIPE NCC service region only concern address space with the status Allocated PA or Assigned PI. Because legacy space is not subject to the RIPE transfer policies, the RIPE NCC cannot track or publish transfers in this area. Inter-RIR transfers are still too small in number to see any trend. In terms of the number of transferred IPv4 addresses, statistics are dominated by a single transaction, eight /14s transferred from ARIN to the RIPE NCC in December 2015.

Linus Torvalds at TED

Sat, 21 May 2016 08:20:06 -0400

If you have not seen the interview of Linus Torvalds at TED from February, I highly recommend it. He’s changed the face of modern computing twice. His code runs in billions of devices (you might be reading this on one right now), and millions of software developers around the world manage their code thanks to his creation of Git. And yet, he’s a humble, gracious and down-to-earth man.

Poll: What's the First Thing You Look for in an Editor?

Sun, 15 May 2016 13:22:06 -0400

What's the first thing you look for in an editor? There is an obvious right answer.
— Andy Newton (@MoreThanABitOff) May 15, 2016

Not Just an Annoying Voice but also a Bad Taste in Men

Sun, 15 May 2016 10:22:06 -0400

This fracas resurfaced in 2014 when Ayyadurai married Fran Drescher.

from an article at Ars Technica about Shiva Ayyadurai suing Gawker Media for libel.

Gawker’s Gizmodo published several articles (1) (2) on Shiva Ayyadurai and the result:

Months after these articles appeared, Boston Magazine reported that as a result, Ayyadurai’s “speaking engagements have been canceled, the funding for his e-mail lab has evaporated, and his contract to lecture in MIT’s bioengineering department has been revoked.”

And here’s the chortle factor:

Harder further argues that Tomlinson’s work (his RFC #561 describes “network mail” in 1971) is too ancient to be considered proper e-mail.

Reality is HARD!

My Guest Post on APNICs Blog

Sun, 15 May 2016 09:22:06 -0400

What does Port 43 WHOIS protocol have to do with an 8 track tape? Answers herehttps://t.co/h0bgJIEwpM pic.twitter.com/Yrh716I2bt
— APNIC (@apnic) May 12, 2016

.CZ Is First RDAP TLD (UPDATE)

Fri, 13 May 2016 13:22:06 -0400

I just noticed that IANA has updated their bootstrap files, and NIC.CZ, the Czech TLD, is in there.

UPDATE : Because the RDAP IANA files have been updated soooooooo infrequently, it has gone unnoticed that NicInfo was not bootstrapping properly. That has now been fixed in 1.1.1.

Big Changes for JCR in the Works

Thu, 12 May 2016 16:22:06 -0400

We published JCR -06 back in March and since we’ve received a lot of feedback. And that feedback has resulted in some fairly significant, and unfortunately backwards incompatible, syntax changes – all for the better.

The big thing to take away from these changes is that valid JSON is now valid JCR. To some extent this was true with -06, but things like this were not:

[ "a", [ 1, 2, 3 ], "b" ]

Well, now it is (or at least will be in -07). But in doing this, we’ve had to move the repetition syntax to the end of the definitions. Whereas we use to do this:

[ :"a", [ 3 :integer ], :"b" ]

repetitions are now preceded with some syntactic sugar, the @, and moved to the end:

[ "a", [ integer @3 ], "b" ]

For a refresher on the concept of syntactic sugar:

In computer science, syntactic sugar is syntax within a programming language t hat is designed to make things easier to read or to express. It makes the language “sweeter” for human use: things can be expressed more clearly, more concisely, or in an alternative style that some may prefer.

Or in other words, the computer doesn’t need it. A JCR parser doesn’t need the @, but we’ve added it there so repetitions don’t get lost and appear to have a meaning more obvious than, “look, they forgot a comma”.

Syntactic sugar is one of those eye-of-the-beholder things, as I’ve learned through this process. And so… we’ve added more.

As you may recall, in JCR rules can be broken out and referenced by name. An example:

[ "a", ints, "b" ]

ints [ integer @3 ]

To the casual reader, that might seem confusing. Is ints a JCR thing? To make it clearer, we’ve decided that rule names must be immediately preceded by the $ character and the = character is to be used when naming rules. Thus:

[ "a", $ints, "b" ]

$ints = [ integer @3 ]

Now those aren’t all the changes we are working on, but those are the big ones. More to come…

XML is Hard!

Thu, 12 May 2016 11:22:06 -0400

This thread on the REGOPS mailing list just makes me sad.