Archive for May, 2007

Gen is Eclipse’d !

Having looked at the new release of the Diagram Trace Utility (DTU) in Gen 7.6 and discovering that it’s based around Eclipse is a good thing. This means that the folks at CA are getting their act together and realising that there is a community of developers out there who know Eclipse and can use this knowledge to transition to Gen easily.
Of course, the rest of Gen isnt an Eclipse application yet, but who knows ? It could be with r8

When r8 arrives, we SHOULD see some improvements to the user interface (if the guys that develop the tool listen to the users – and we have proof that they do) and hopefully it will be a VAST improvement on what we have been living with for years. Not that I am being critical – no – far from it – the pointandclick interface generating syntax-error free code has been around for many years in Gen and was a masterstroke – but resting on your laurels too long means those laurels become grey and withered. Time to move on to another innovative interface or (better yet) get on the Eclipse bandwagon and use that “made with Eclipse” notion to lure developers, testers and architects to the Gen world.

Gen and deployment

Whilst at CA-World in April, it came to my attention that many of our American friends (and in fact some organisations in Australasia) seem to prefer a deployment platform of a mainframe. Most of the Gen organisations that I am aware of in the UK seem to prefer the distributed (i.e. Unix) environment.

 Call me cynical if you like, but isn’t the mainframe dead ?

 Well – if you listen to the people that talk about SOAs and Java and stuff then yes – it is, but if you listen to the realists, then what they want is a platform which is reliable, flexible and powerful enough to take anything that you can throw at it.

 Indeed, some of the people I spoke to in Vegas have systems that would dwarf what most of us have dealt with – running on fairly new mainframes. In fact, those mainframes run a combination of z/OS and Linux and (gasp) participate in SOAs quite happily.

 So – you tell me then – how is it that a toolset like Gen that has been around for ages, deployed on a platform that has been around for as long as computing has been around (..the “main frame” of the computer) – can expose services and consume services ??

Its not java – its not eclipse, its not coffee-related – it doesnt have a name which seems to come from a Tom and Jerry cartoon – it just works!!

 Don’t get me wrong – I dont have a downer on modern toolkits its just that they seem to be “the next best thing” ALL the time – when in fact, the EXISTING best thing is staring us all in the face – existing toolsets (Gen or otherwise) and mainframes.

The site in which I work doesn’t deploy Gen on a mainframe, but I wouldn’t be adverse to using one – after all – these days we want our “grid computing” (read MIPS on demand), SOAs (read expose existing CICS transaction thru an interface) , thin client (read dumb terminal or telnet emulator) and so on.

The point I am trying to make here is the NEXT BIG THING is already here and has been since the dawn of the computing era – it just has a new name now – think of the children’s story of the Emperor’s new clothes  – just the same but different !

 Gen can be dressed up in 1970′s denim and leather (batch programs) , 1980′s New Romantics (dumb terminals) , 1990′s Hip-Hop or Acid House (Client/Server) or 2000′s manufactured “Boy Bands” (SOAs).

Which will you chose today ?

Gen and SOA

I’ve been looking at the ways in which people search the net for gen-related topics, and (of the small number of requests there are) lots of people seem to be searching for “Gen” and “SOA” in the same breath. To me, this means that either a) people want to know if Gen can “do” SOA, or b) people KNOW that Gen “does” SOA but “How does it do SOA” ?

I’m not really sure, since there are a large number of significant Gen sites around the world, and I presume that many of them already “do” SOA.

The fact that an SOA is an architecture, and has standards, and as long as a product/toolset that you want to come to the SOA party can conform to those standards, then that product/toolset should be allowed an invitation to said party !

Gen has had an invitation since before SOA was called SOA, but people probably don’t know that. The fact that it can consume and expose web services, generate XML interfaces and EJBs (buzzword overload) means that, with the correctly architected infrastructure, it can not only come to the party but bring lots of presents, too.

The presents that Gen brings in my opinion (shameless plug for comments) are stability, robustness, simplicity (complexity is fine, but do you really need it all the time ?) and technology independance

But hang on – technology independance – isnt that what SOAs are supposed to bring to an enterprise? – standards are only really required when you want one vendor’s product to talk to another vendor’s without buying some expensive, proprietry interface software or hardware. Bringing technology independance to a technology-indepentant party is a particular strength of Gen – there exists a whole layer of independance which should not be ignored – that of time-independance.

“Time-independance” is an offshoot of technology independance. This means that the code that you produce today is independant of the time in which it was produced – meaning that, save for business requirement changes, which affect all systems, the code that was produced years ago is STILL relevant today, and, more importantly, deployable on today’s platforms. So, to say that Gen produces technology-independant systems, while correct, does not emphasise that the systems it produces stand the test of time, despite (or because of….) platform and technology change.

21st Century Gen

Following on from last week’s thoughts – a more general ramble on the “appearance” of tools of today and how CA-Gen stacks up.

We all know how dated the user interface of Gen looks – but does that really affect its productivity ? It certainly affects its image, which pobably in turn affects its’ take-up. Users these days seem to only like and adopt a product if it has some fancy name and is coffee-related .

In point of fact, this may have been tried a number of years ago – trying to make a product “Cool” by calling it “Cool” didn’t in my opinion work-  using a name like “Advantage…whatever” didn’t work either. Simply because a product has a whizz-bang name doesn’t make it acceptable and attractive to people to consider taking it on.

Yes – all of us within the Gen community know and love Gen for what it can do, but how to promote that message outside the community to widen it and sow the seed that will grow into a large Gen-based tree is difficult to conceive.

Perhaps (as I said last time) focussing on productivity, power, robustness and flexibility is the way to go here rather than focussing on marketing spiel ??

So, keeping the name at CA-Gen for more than a few weeks will probably help – a change of name hardly does anyone any good, least of all a software product !

Where should Gen go ?

Looking at the users of Gen worldwide, most of them seem to be going the SOA route.
Whilst this is not inherently a problem, and whilst Gen can participate fully in the SOA world, just as easily as it could in the batch world, the mainframe world and the client/server world, what part SHOULD Gen be playing in this world?

Should Gen be a service consumer as well as a service provider ?

I believe that Gen should be marketed primarily as a back-end development tool – a service provider - with front ends done by other tools.

Gen’s strengths are in high-volume, reliable, sturdy and stable service provision – let’s push that ! 

Let’s face it, the DreamWeavers of this world have got the user interface on the web sewn up. What we don’t need is YAWIT – Yet Another Web Interface Tool.

What we need is a tool that can provide the services at the back end. In these days of corporate databases and large,complex datasets, we need a tool that “just works”. Gen IS that tool. Yes – it can “do Web” but do we want it to “do the whole shebang ??”

I firmly believe that the place that Gen should be is at the back end – providing the services for the fluffy tools to hook into to provide the user experiences. In fact, if Gen were a pure service-creation tool, then those services could be used by a 3270 interface, a Web interface and a client/service interface – now there’s an idea !