Jeff Panici

Why is it every year that goes by is harder not easier?

Come on everyone, we're programmers; can't we make some of this stuff any easier? EJB is certainly not easy.

David, I've been around. When the whole TopMind, Costin, god knows who battle erupted I decided I had better things to do with my time. I said to myself a while back I'd keep an eye on things and come back slowly when thing appeared to quiet down. It may be close to that time.--JeffPanici

You got me curious about TOP contributions which I have not been following. Assuming the views he expressed are controversial, do you find him frequently involved in "provocative behavior"? He is colocated in a community that is strongly favoring OO so I would imagine there is a "natural impedence mismatch" -- dl

TopMind refuses to accept that others might not agree with his views. While some of his contributions are valid, his repetition and intrusion in an inordinately large number of topics here is a waste. I'm sure others would say the same about me, if given the chance. Regardless, I tire easily of rehashing the same old topics. My general feeling today is that the state of software development is near abysmal. The tools and languages are sufficient, at best. I've been heavily using .NET and C# and must say that, while much better than anything I was using 10 years ago (with the exception of Smalltalk, of course), one begins to see the cracks in the facade quickly. Part of me thinks this is a more fundemental thing, in that the operating systems we use today from Windows to Linux are fundamentally outdated and we keep banging our heads against the brick wall as we grow. We're busy building on a broken foundation and constantly trying to fill in the cracks. Hence, to some degree, I think time spent arguing about which hole to patch first is fruitless and just a bit depressing.--JeffPanici

I've been programming since I was about five (hey some kinds learn the piano -- I learned how to use a soldering iron and a keyboard). My first few programs were written in z80 assembler, then BASIC, then Pascal for a long while, PROMOL, and a whole host of other bizarre languages. Fourth was truly awesome, if somewhat odd. In my formative years (I was a kid), I mostly wrote game software, some of it was actually popular in its time. The TRS-80, Commodore 64/128, Commodore Amiga, Apple ][ and such were my learning grounds.

I have to say I really loved Modula-2 on the Amiga. Ada95 is the cool modern version. Of course, I've done what seems like decades of C/C++ coding. Alas, I find myself writing a lot of Java these days. Worse, while Java itself has its fair share of problems, many projects are treating Java as a procedural language, or writing EnterpriseJavaBeans. I'm sorry, those silly beans are just pure evil and they send us right back to the dark ages.

I'm a consultant, striving to put right what once went wrong and hoping that the next leap ... will be the leap home.

I have come to understand that there are many programmers who only think they're programmers. huh?

Many talkers and very few doers.

Hi Jeff, I like your ideas. I started at 10 not at 5 with a Z80, on to the 6502 (Apple][), 8088 and 386, Smalltalk and finally Java. I'm trying to make things easier, take a look: PrevalenceLayer. -- KlausWuestefeld

I've recently been doing a lot of .NET work. I think the initial updraft in interest for .NET is starting right about now. I'm curious how deeply .NET will work its way into the market. At this point, C# (and #Smalltalk!!!) is a good alternative to Java.

On TheyWillNotListenResolution (from 2003) you wrote: We eventually convinced the client to use Oracle running under Solaris. This took some pushing by IBM as well; they finally admitted the mainframe wasn't going to do what we wanted. We created our own in-memory object cache fronted by some simple RMI (Remote Method Invocation) objects to eliminate repeated queries against the mainframe DB2 database.

I know its almost ancient history now, but can you elaborate -- how much of the improvement was due to the move to Oracle/Solaris (vs DB2/zOS), how much was due to the caching, and how much was due to (the implied) dropping of TopLink? Any idea how fast it could have been on zOS if differently architected? (e.g. did you try the raw SQL queries for a typical query screen from a COBOL program to compare speed).

''Yikes, you're right, this is almost ancient history. I'll try my best to dig through the ol' memory banks. I'll take these slightly out-of-order:

1. The centralized object cache was significant and, in hindsight, probably represented the most gain of any one change in the architecture. Keep in mind we literally took approximately 125 (I think this number grew over time) tables and brought them all into memory as objects, addressable by complex dictionary keys. We designed some intelligent proxy classes that made it seamless to request these object graphs locally or remotely. The proxies (clients) of the remote cache *also* cached instances as they were serialized over the wire so the cost of roundtrips reduced over any one runtime session (defined by a startup and shutdown of the application server). The cache refresh strategy was easy: Restart twice a day when product, customer updates were done (in batch) on the mainframe. I forget the exact times but it was something like 5:00am and 3:00pm. The restart time was less than five minutes and I think that was improved over time as well.

I need to finish this off a bit later. WIP -- JeffPanici''

CategoryHomePage