Well. I was at the Systemideveloper.com spring conference last week and managed to let my voice be heard during George Farr’s speech to the masses – us RPG programmers. First let me say that I definitely respect George and appreciate many of the things his team does that helps my development efforts. I don’t think I could code RPG in SEU anymore and have fallen in love with WDSC’s RSE and LPEX editor.
With that said I am getting red flags with the things I hear coming out of George’s mouth. Now, I am not one to conjecture, but I can’t help myself in reading between the lines of what George is saying. What is George saying? He is playing his horn to the tune of platform independence and playing it quite happily. Platform independence inherently means Java in my mind (he also mentioned EGL), and that doesn’t bode well for us RPG programmers if that is where IBM would like to see us in 5 to 8 years. I have been a Java programmer for a number of years now (roughly 5) and while I would not consider myself an expert in the language, I have definitely been dating long enough to know that marriage in the town of iSeries is not in the future (if you catch my drift).
Java IS a great language for MANY scenarios (i.e. frameworks, running on printers and the like), but of those it is NOT a replacement business logic language for the iSeries. You see, George gets his ear talked off by vendors that would LOVE to have the iSeries platform to run their Java/EGL web apps. After all, we do have one of the most solid platforms out there. But an iSeries isn’t always easy to sell when you can get an “equivalent” hardware machine for half the price. So that is why vendors aren’t lining up to write next gen apps in RPG. They want to be able to give their customer sub-par hardware and OS (i.e. Wintel) when the customer says the iSeries costs too much. What George doesn’t hear are all of us RPGers that are in the trenches making businesses run and don’t always have time for the scholarly task of learning a scholarly language. We instead are writing code, very fast, so our companies can stay afloat and make money. Should we be slapped around for not getting in the face of George and telling him what we need vs. him giving us what he thinks we want – probably. George’s challenge is to embrace one of the most dedicated programming communities and figure out what makes them tick – what could we use that will take us into the next 10 years of programming. We don’t ask for much, just that we don’t have to learn Java or us it during our runtimes. We simply want four things…
- a native web framework, much like JavaServer Faces, but that has nothing to do with Java. When a user clicks a link or button I want to be able to “inject” that right into an RPG sub procedure in my RPG CGI program. CGIDEV2 was great 5 years ago when everyone on the iSeries was learning about HTML, but it just doesn’t cut it for next generation web apps.
- a native way to build thick client GUI applications. Use the Smart Client concept so we don’t end up with Visualage RPG. Make all the RPG code run on the iSeries so we don’t run behind in compiler releases. Oh, and WDSC would contain a no additional cost drag’n’drop GUI builder.
- language modifications that get us completely into free form. Note that D, H, and F specs aren’t so bad, but the change is mostly so we can market our language to the next generation of programmers (disclaimer – I am 27 years old as I write this).
- Rename the language. I am sick of having a Cannondale bike with Huffy stickers plastered all over it. RPG is hardly Report Program Generator today. Interesting how IBM has rename diarrhea for the platform and OS, but RPG has hardly changed names in the past 12 years.
If IBM would get their druthers straight they would realize that platform dependence is an incredibly great thing (for their customers) because it provides tight integration (i.e. RPG/DB2/OS400) allowing fast development of applications, and it inherently keeps customers on the platform which means lots of hardware sales for IBM. Don’t think it can work? Ever heard of a company named Microsoft? How many Microsoft products run natively on Linux? On Unix? On OS xyz? They don’t, and Microsoft gains MANY competitive advantages because of it. (yes I know about mono – hardly counts as Microsoft dropped their support of it a year after they started the initiative) IBM doesn’t even have to pioneer this, they simply have to copy the most successful marketing technology company in the world.
With all that said here is the press release that started this whole thread on SystemiNetwork. See the last couple paragraphs for mention of my name – Aaron Bartell.
[tags]rpg, java, egl, wdsc, ibm, System i, iSeries, os/400[/tags]
As an old boss said to me, and I reiterate here, companies buy the platform base on the software, not the software based on the platform.
I agree with Aaron I would like to see a drag on drop GUI approach, seems like in intuitive way to move us RPG’ers along.
Good point Kevin. I am finding not only is it people buying based on software/applications but almost just as much as the development tools. People move to the proprietary windows platform many times because Microsoft has such great development tools/IDE’s. Very interesting that programmers at the lowest level of the IT food chain, are making decisions based on which language they like the best. Thanks for your comments Kevin.
Hi Aaron
Microsoft has basically given away its development environments, compared to what IBM is doing. And guess where a lot of development is happening.
There are things internally at IBM that probably mitigate against making WDSC too much of a loss leader. But it should be looked at more seriously.
Later
I disagree with your point about platform dependence.
This strategy is working for Microsoft, because they have broken the critical barrier on market penetrance a long time ago.
Additionally, Microsofts software and the hardware needed to run Microsoft’s software is pretty cheap, compared to a System i.
And one more point – you will always need a Windows server, as long as your clients are Windows based.
And IBM’s pricing for the System i and it’s software is horrendous.
>This strategy is working for Microsoft, because they have broken the critical barrier on market penetrance a long time ago.
I agree that M$ doesn’t need to do much in the way of working with other platforms simply because most desktops ARE M$’s. Still doesn’t change my stance though, because look at all of the intimate integration they have been able to obtain by not having to work with parties outside of themselves. The iSeries has very similar advantages with i5OS, DB2 and ILE, but that is being lost with IBM thinking that platform independence is a must.
>Additionally, Microsoft’s software and the hardware needed to run Microsofts software is pretty cheap, compared to a System i.
I can’t argue the price point, but I think they live at different levels. The System i is an enterprise server (hardware and software) whereas Microsoft is not (yes I know they are trying, but I don’t believe they are even close yet). SO in the end I would EXPECT to pay more for an iSeries because I am getting so much more. The real problem then becomes how does one get introduced to the platform if they can’t buy one and run it in their home.
>And one more point – you will always need a Windows server, as long as your clients are Windows based.
Not sure I follow you here. I work in a remote office (my home) and do not have any Windows servers. I DO have Windows desktops though. But I am connecting to many different servers throughout the day including iSeries, Linux, etc. Up until earlier this year we didn’t have ANY windows servers in our environment because we hosted our web stuff on Debian. Again, I am guessing I don’t understand what you are saying.
>And IBMs pricing for the System i and its software is horrendous.
Here’s a link to an entry level system that my company recently purchased: http://mowyourlawn.com/blog/linkto/systemiconfig.pdf ($12,904)
I agree that it isn’t cheap, but again we get back to the discussion of an enterprise server vs. a desktop machine. I wouldn’t run an enterprise on windows “servers” simply because history has shown they simply aren’t made for that. Again, a costly machine is crippling new developers from coming onto the iSeries.
Thanks for your comments Lukas,
Aaron Bartell
http://mowyourlawn.com
The problem here is longevity. Most small businesses can get very attractive pricing for Microsoft’s Small Business Server. And when they grow, and grow, and become an enterprise, they will already have a mostly-Microsoft Infrastructure, and never even consider the System i. (There are exceptions here, e.G. if a Vendor offers are very interesting software product only on the System i, but the chance for this is rather slim)
Of course you don’t need a windows server at home, but if you’re running a company with more than a few computers, you will need to manage deployment of security patches, program updates, configuration, etc.
The only, currently existing alternative to Windows Clients are Thin Clients, accessing a Windows Server. IBM doesn’t have any offering in that field.
No matter how good the System i currently is, it can’t compete with Microsoft on the Client market. This ensures the existence of Windows in that segment.
Linux has big market share in the infrastructure environment, even in bigger companies, but only by displacing existing Unix machines.
I can’t really talk about the enterprise market, because i’m working in the small business market.
Another important thing you might want to consider.
Im currently aged 22, and so far i haven’t found anyone else doing technical work on the System i in switzerland in my age bracket. The reason for this is simple.
When i was younger, i experimented and learned with windows server at home, using an illegal copy. This helped me to build knowledge on that platform, and this is currently to 60-80% my main job.
OTOH, i only started working on the System i because of unusual circumstances, i would have never considered starting in this field, because it is simply to unattractive.
I don’t have a problem with 5250 for administration tasks, a vt100 for my linux servers isn’t much different, the problem is that no one knows the platform, and using 5250 for end users is no longer an option.
A baseline 520 also doesn’t cut it for any type of java work, and i don’t see why i should use a castrated 1.9ghz cpu for running java when i can get 4 uncastrated 2.4 Ghz cores at the same price.
Every hand is a winner and every hand is a loser. No matter how good RPG on the iSeries is, if the market goes to Microsoft and Java them RPG will sink into oblivion. That RPG is still here shows what it is capable of.
C# and Java are yesterday’s paradigm. Tomorrow’s winner is the Rails framework and it’s clones. If you’ve been paying close attention, you know that Microsoft is soon going to introduce Ruby on Rails into Visual Studio. Katy, bar the door! The meta-programming paradigm is going to replace line-by-line coding as we’ve done it.
If George can create a Rails-like framework around RPG then RPG will stay in the game. Otherwise, cue Don Merideth.