Just noticed that the sale of Budweiser was approved to the Belgian company InBev according to CNN reports. Perhaps they'll make it taste more like Belgian beer... you know, not crap.
Actually, add 5% more alcohol and add some Gamay flavor and you got a North American winner on your hands.
Monday, July 14, 2008
Thursday, July 10, 2008
The debate rages while the underground bubbles...
Very interesting little debate going on over on Steve Vinoski's blog on RPC vs. REST. I'm personally a big fan of the REST style these days for all the typical reasons: caching, looser coupling, standardized interfaces, uniform access, etc. However I would say that I'm not a "REST is best, SOA is DOA" - I think SOA does indeed serve very nicely at providing a unifying mechanism around various RPC and messaging systems allowing integration and interoperability. I'm not sure that REST as implementated today would be as well suited but I'm open to opinion. I also very much agree with the comment made by Dan Diephouse, namely that most companies have a ton a decent Java programmers comfortable with their language and prefer the way RPC allows them to express their ideas 'natively'. I was primarily a C++ programmer so I've always seen Java as the lazy mans language anyway so I'm not surprised they default to RPC and stay there. :) But Java programmers are a commodity so I can understand the decision to allow them to work in their comfort zone.
(I love when zealots argue btw because you can really dig into an issue quickly. The RPC vs MOM wars of old did leave a wrath of education in its wake, by the way. :) )
Where Steve really nailed it though is on the fact that there will soon be required a fundamental shift in the lingua franca of development. Applications designed around the idea of shared state and synchronization will become far too unwieldy to manage - deadlocks, race conditions, nasty overwrites, contention - all much harder to fix, and scale! Cores will soon scale to the hundreds and who knows what kind of shared hardware interconnects will come from the datacenter shift. Erlang might indeed be the next wave, who knows? But I do know that nothing last forever and my guess is that Java has peaked at 5.0 and by 8.0 - 9.0 timeframe we'll see a revolution occuring. It's still underground but its bubbling.
Java has pulled in annotations which Lisp had 20 years ago if I'm not mistaken as well as templates which I vaguely seem to recall were frowned upon as too abstract and difficult for the language at inception. Pretty soon they'll have unsigned ints, overloaded operators, and all the Lisp and C++ features... perhaps inline assembly, and even multiple inheritance! And the commodity developers will be so jaded they'll turn to some simple language like Erlang.
(I love when zealots argue btw because you can really dig into an issue quickly. The RPC vs MOM wars of old did leave a wrath of education in its wake, by the way. :) )
Where Steve really nailed it though is on the fact that there will soon be required a fundamental shift in the lingua franca of development. Applications designed around the idea of shared state and synchronization will become far too unwieldy to manage - deadlocks, race conditions, nasty overwrites, contention - all much harder to fix, and scale! Cores will soon scale to the hundreds and who knows what kind of shared hardware interconnects will come from the datacenter shift. Erlang might indeed be the next wave, who knows? But I do know that nothing last forever and my guess is that Java has peaked at 5.0 and by 8.0 - 9.0 timeframe we'll see a revolution occuring. It's still underground but its bubbling.
Java has pulled in annotations which Lisp had 20 years ago if I'm not mistaken as well as templates which I vaguely seem to recall were frowned upon as too abstract and difficult for the language at inception. Pretty soon they'll have unsigned ints, overloaded operators, and all the Lisp and C++ features... perhaps inline assembly, and even multiple inheritance! And the commodity developers will be so jaded they'll turn to some simple language like Erlang.
Tuesday, July 08, 2008
G8 bullshit and an idiot Canadian leader
To say that I'm absolutely disgusted with the G8 decision to half global emissions by 2050 would be the understatement of the week. That kind of mark is so far off it's foolish. By 2050 most of the these current leaders will be long dead (and good riddance - they're all corporate shills anyway), so it's an easy cop-out for them. Where are the milestones and targets? Do we expect to radically change in 2049 in order to meet deadlines?
2007 was the first year the Northwest Passage was clear to sail. 2008 will be the first year the North Pole is open water. So let's half emissions by 2050 but let's not set shorter term goals to address imminent shorter term threats? Wow. Idiocy rules I guess. It's so unfortunate that big talkers and cocky egos run companies and governments...
At least there are some folk serious about this however. Mr. Pickens announcement today at least reassures me that there are some sensible, do-good humans bucking the human-virus condition.
Canada is a country I'm very proud of but the current leadership seriously needs to be put to pasture - in Guantanamo! Praising the 2050 deadline, standing by and doing nothing while millions of Africans are murdered, brown-nosing our way into GWB's ass and going off to Afghanistan, and upping defense spending by what, $50b over the next 10 years? That's not the country values I'm proud of - that's the values of an arrogant, stupid, ego-maniacal leader. America unfortunately doesn't have a monopoly on such leaders.
Dion is angry today.
2007 was the first year the Northwest Passage was clear to sail. 2008 will be the first year the North Pole is open water. So let's half emissions by 2050 but let's not set shorter term goals to address imminent shorter term threats? Wow. Idiocy rules I guess. It's so unfortunate that big talkers and cocky egos run companies and governments...
At least there are some folk serious about this however. Mr. Pickens announcement today at least reassures me that there are some sensible, do-good humans bucking the human-virus condition.
Canada is a country I'm very proud of but the current leadership seriously needs to be put to pasture - in Guantanamo! Praising the 2050 deadline, standing by and doing nothing while millions of Africans are murdered, brown-nosing our way into GWB's ass and going off to Afghanistan, and upping defense spending by what, $50b over the next 10 years? That's not the country values I'm proud of - that's the values of an arrogant, stupid, ego-maniacal leader. America unfortunately doesn't have a monopoly on such leaders.
Dion is angry today.
Saturday, June 28, 2008
I'm a sun worshipper
At the end of May I just down to MA, USA from my life-long home of Newfoundland. Already I've received more sun than I would have for the entire summer in NL, and that's not a joke. My skin is darker - though I'm still pasty white - vitamin D levels have shot up, and those pesky riskets are nearly gone. In short, I've become a sun worshipper.
Besides increased sun exposure, another thing I've noticed about the US is the permeation of religion. Churches are everywhere and bumper stickers proclaim that the occupants have found Jesus. Myself, I could care less for religion though I'm certainly tolerant of people's beliefs - my girlfriend is Muslim and we've managed to get along. However, in light of my new pagan god the Sun and the passing of my favorite comedian George Carlin, here's a link I'm sure some of you will enjoy and/or remember. For those of you strictly religions, avert your eyes and ears.
Religion is...
Besides increased sun exposure, another thing I've noticed about the US is the permeation of religion. Churches are everywhere and bumper stickers proclaim that the occupants have found Jesus. Myself, I could care less for religion though I'm certainly tolerant of people's beliefs - my girlfriend is Muslim and we've managed to get along. However, in light of my new pagan god the Sun and the passing of my favorite comedian George Carlin, here's a link I'm sure some of you will enjoy and/or remember. For those of you strictly religions, avert your eyes and ears.
Religion is...
Friday, June 27, 2008
The software long-tail?
I assume that everyone is well familiar with Chris Anderson's best seller The Long Tail, or at least the concept of the long tail. Slashdot, one of my favorite sites on the net of course, referenced a very interesting article from the Harvard Business Review disputing the theory of the long tail effect. The long tail effect made so much sense on the surface that I never really gave it a second though - I just bought into it lock stock & barrel. But after digesting Anita Elberse's article for a while it got me to thinking about the long tail effect on enterprise software.
With a tangible good such as a CD, DVD, book or anything occupying space and incurring shipping costs, it's hard to argue with the points Anita makes. Considering something like digital music however, I tend to agree with Chris's position on the long tail. Since both shipping and storage costs are virtually nil - hard disk storage and internet connectivity being readily available - its not hard to turn a profit.
One would think then that enterprise software, or most software for that matter, would reap the same benefits as digital music since storage and replication are effectively cost void. Unfortunately software is much more dynamic than music, requiring constant nurturing and TLC. One can easily keep a digital copy of Tom Wait's Raindogs stuffed away on a disk somewhere until it needs to be put on the wire - it doesn't need to change and likely won't need to do so unless a new digital music encoding format is required, and even then a transformation is trivial and cheap. Software on the other hand requires constant porting to new OSes, JVMs, compilers in order to remain useful and modern. Patches and fixes are often required, and interoperability (at least in the enterprise world) will continue to require leg-work and incur cost. How many copies of Word Perfect 5.1 for DOS were sold last year?
The long tail effect might be less prevalent than one might imagine. Make sure you analyze the cost vs. potential - that tail might be bobbed.
With a tangible good such as a CD, DVD, book or anything occupying space and incurring shipping costs, it's hard to argue with the points Anita makes. Considering something like digital music however, I tend to agree with Chris's position on the long tail. Since both shipping and storage costs are virtually nil - hard disk storage and internet connectivity being readily available - its not hard to turn a profit.
One would think then that enterprise software, or most software for that matter, would reap the same benefits as digital music since storage and replication are effectively cost void. Unfortunately software is much more dynamic than music, requiring constant nurturing and TLC. One can easily keep a digital copy of Tom Wait's Raindogs stuffed away on a disk somewhere until it needs to be put on the wire - it doesn't need to change and likely won't need to do so unless a new digital music encoding format is required, and even then a transformation is trivial and cheap. Software on the other hand requires constant porting to new OSes, JVMs, compilers in order to remain useful and modern. Patches and fixes are often required, and interoperability (at least in the enterprise world) will continue to require leg-work and incur cost. How many copies of Word Perfect 5.1 for DOS were sold last year?
The long tail effect might be less prevalent than one might imagine. Make sure you analyze the cost vs. potential - that tail might be bobbed.
Subscribe to:
Posts (Atom)
