Popularity

mgb is attracting the wrong sort of people (well, seemingly uncharming and illiterate) but maybe that's what you get for talking about beer.

I have only one thing to say
about this [via James Gosling] and I'm not going to censor it. Fuckers. The full first-hand story is here.

Solve these problems:
When was the last time you were at a meeting and were trying to reschedule the next one? Each person calls out suggestions based on their own calendar and everyone else checks that date, someone inevitably rejects it and the cycles starts over. Soon your come to the conclusion that there is no perfect date so a compromise is sought. Why isn't this sorted out by now? Why don't we have diaries which can negociate with each other? We don't even have the ability to discover which days have the least collisions?

One of the reason that this problems hasn't been satisfactorily solved is that everyone is using different devices, different operating systems and different PIMs. The information is locked away in the applications and must be exported to make use of it. Information doesn't flow, its pumped. It involves the expediture of energy to move it from one place to another. Perhaps database-style filesystems will help, perhaps common, well-known information stores like the address book on the Palm are the answer or maybe the semantic web will save us.

These tasks really shouldn't be so hard.

JXTA or Jabber
For mobile messaging-style applications which would you choose, Jabber or JXTA for J2ME (JXME)?

Jabber has lots of interesting features for instant messaging like "presence" (which JXTA is only just getting around to) and even "mood indicators". But overall, JXTA has a nicer architecture which would seem more appropriate to building applications on top of. JXME also has the advantage of being smaller (~10k[pdf] vs 50k?). And apparently Sun are beginning to think of JXTA as something other than a toy research project. On the other hand, Jabber clients already exist for J2ME whereas I'm not aware of anything based on JXME.

But maybe there is a compromise to be made: Jabber over JXTA? Sure, this might negate some of the lightweight-ness of JXME and Jabber purists might argue against using Jabber XML stanzas outside of a TCP stream. Oh, and JXME would have to be improved in some areas and would require MIDP 2.0 (I have a draft post on this which I'll publish soon). But it could be worth investigating.

p.s. I'm not actually thinking of building an IM client, just something like it. but so much cooler.

The Mailinator
I've seen this mentioned in a few places, but first by Joel. The idea is that you can create a transient insecure email address for those pesky web forms that require a valid email address - thus avoiding spam on your real email account. There's no setup, no passwords, no security and no hassle. Any email that arrives gets dropped into the appropriate inbox for a few hours where anyone can read it. From the Mailinator FAQ:

This sounds pretty insecure. What if I send important emails with sensitive super-secret information in them to mailinator?
Then you are a stupid-head. That isn't what this is for.

I like this attitude to security - if something is meant to be insecure just be upfront about it.

My only concern is that eventually the site will be receiving so much spam that it'll shutdown. Ideally the addresses should all be once-offs so that spam can easily be identified and ignored/reported.

Attention FreeRoller users and the people who link to them
I've just got my daily dose of Erik and of the five links that caught my interest 3 were hosted at FreeRoller. Of course since the FreeRoller site has the worst uptime of any blog host I've yet found I couldn't read the following posts:
PalmOS moves more 2 J2ME.
Lucene and Keywords: Disappointing.
SweetReader project open at SourceForge.

If you're thinking of a free blog host, please please find out about their uptime and consider how badly you want to annoy your readership. It's rather unfortunate that many of Erik's links are to FreeRoller sites - they obviously have something interesting to say and FreeRoller doesn't do them justice...

update : even though FreeRoller blogs are back up at 5:30pm, they're still dog-slow.

Zildjian
Alison has a cool picture of a Zildjian cymbal. Dunno why you'd be interested but I like it.

research
The latest issue of ERCIM seems to have been picked up by everyone. It's odd because this publication has been around for years and I only discovered it 6 months ago. Now everyone's pointing to it. I quite like the concept of this journal: its structured as a series of short 1-2page abstracts around a particular theme of research. Each abstract describes an ongoing project or published paper which is then referenced so you can get a good feel for a topic and be able to follow up the papers in more detail. Oh, and being free in digital and paper editions is very nice.

Unfortunately, I was quite depressed by the content of the recent issue on "Applications and Service Platforms for the Mobile User". I couldn't spot one architecture which was likely to make as a commercial entity which is fine if you are only building that platform to investigate another aspect (say, new types of applications or new styles of interaction). Unfortunately, these platforms seemed to exist in a bubble, separated from the rest of the world and, whilst interesting in their own right, I didn't understand why much of the research was being done. Where was it going? What results could I actively use if I was to start creating a mobile application today? The applications didn't exactly spark my imagination either: mBlog was probably the most interesting but there was so much more that could have been done. For example, where was the ability to get location-aware RSS streams, discover local bloggers or handle the whole blogging without a centralised pre- and post-processor? And these were just to my off-the-top-of-my-head questions.

This attitude is partly due to my recent annoyances at inventing things 4 years ago which have still not materialised. Or indeed, reading papers which were written in 1995 but whose ideas have yet to be implemented. I'm having a crisis of belief in the relevance of my research ("Why O' Why am I doing this?") and I'm beginning to over apply those thoughts to research in general.

There needs to be an evaluation of research projects whereby the project doesn't have to use to latest technology or be buzzword-compliant but it does have to go to great lengths to produce "take home" messages for the commercial world. We did X, Y, and Z using A, B and C but the important point of our research is THIS. Another idea is for each group or PhD student to run their own company which is there to commercialise and market the results of the research towards the end of a project. 90-95% of all the companies would fail (which is OK and an expected result) but it would force each project to at least think about the relevance of their research within the wider world. It could even be a paper exercise but making a shell company would add some realism and commitment. The relevance I'm after is not buzzword-compliance. If you're using SOAP in your research, that's wonderful but it is not a cause for celebration nor is it worthy of many journal inches... what have you done that's new and useful? Of course there is always the danger of excessive relevance.

Today we see many innovative products come out of small companies where the "research" was someone tickering away in their spare time before deciding to create the start-up. Blogging is a perfect example of a popular trend which came from the grass roots, not out of academic research. I can't help feeling that there aren't enough bridges between research and development.

Another gripe of mine is that research groups have an unhealthy obsession with building their own platforms (I can count myself here). In the early stages of a technology this can be a good thing because it encourages diversity, competition and innovation. But later on they still continue to produce platforms and architectures without adding anything to the collective knowledge of the field. They produce their own clone of the current implementations because they want to own it (pride), they want to produce it (justify grants) and they want to control it (the it-wasn't-made-here syndrome). I know of a research group which reimplemented their own platform in a different language without significant feature enhancements or creating new knowledge. But it did justify a few students for a year. Thankfully, most agent research groups have started to look beyond building their own platforms and started to look at the truely difficult problems.

Obviously, this whole discussion is particularly biased towards computer science research. Sorry for the rant, I'm in a ranting mood today.

cynical
Did anyone else see the front page of the Motoring section of the Irish Times yesterday? On one side we have an article calling for the reintroduction of the car scrappage scheme which saw cars over 10 years old traded in for new cars and a ?1000 tax rebate. The article made this out to be some great thing for the environment but also noted that new car sales were down 6.5% compared to last year - oh, and did I mention that it was the car industry calling for the car scrappage scheme to be reintroduced? That's right, the car industry was being concerned for the environment.

The other article (which I can only believe is strongly related to the first) is that the used car market is getting stronger and defying the downturn in the used car market.

Does anyone else think that perhaps people can no longer afford new cars and are therefore turning to used cars? And this has absolutely nothing to do with saving the environment...

Ignorant Palm
Russ points to an interview with PalmSource's CEO where he basically complains that the specification of a PDA Profile for Java (within the community-based JCP) did not turn out as they wanted it to, so they've kinda gone cold on the whole Java thing. Interesting business decision when I can be pretty sure that new developers are flocking to the J2ME platform whilst the Palm developer numbers remain static (at best). Surely, wholeheartedly supporting J2ME can only be a good move for Palm? But maybe it isn't J2ME they have a problem with, just Sun. So they'll buy their J2ME implementation from IBM

3.0 or 2.2?
Rusty describes the new features in Eclipse 3.0beta: "Overall this feels more like a 2.2 than a full version shift".

Maybe they're trying to catch up with Netbeans which is currently on v3.5. I've noticed that with some products, the version numbering can often have as much to do with their competitors numbering than with the new features added. Or am I too cynical?

Nice quote to begin the week

"I'm an inventor, and I started looking at long-term trends because an invention has to make sense in the world in which it [is] FINISHED - not the world in which it [was] STARTED." — Ray Kurzweil, via the Harrow Technology Report

Just to let all know...
that Hilary has had a successful tonsil seperation and is now at home resting.

in Ireland...
cows get beaten up on the streets of Dublin (shame on this city), the government are about to ban the postal service from using their shiny new super-trucks and a whale dies near Kinsale. Depressing.

Diego on MIDP
Diego echoes my concerns on the restrictions MIDP applications face. I can understand why we don't have JNI: Many MIDP phones do not have any means to develop for the native operating system (say the Nokia 6310i) - it's only the Symbian phones that could really benefit from a native interface.

What we really need is a JSR for a mandatory API which provides the ability to read (and maybe write but that's not so important) to native databases in a record style format. The API must support access to the phone's phone/address book and should also allow access to the calendar, memos etc if those features are present on the device. I'd envision an API like such:

public interface Database {
public Enumeration getRecords();
// ...
}

public interface Record {
public String getField(int field);
// ...
}

public class AddressBook implements Database {
// field names
public static final void NAME = 1;
public static final void PHONE_NUMBER = 2;
// ...
}

Obviously, there should probably be methods to search the database (if supported) and to retrieve specific records (if supported) but this is an off-the-top-of-my-head sketch of what I'd like to see. Simple, flexible.

update:Stefen wants MIDP application to be able to control phone dialling. The depressing thing is that PersonalJava already had an API to do this and Symbian devices used to support it (still do?).

Java Occam?
Back in Uni we had to write a neural network in Occam which could then be compiled down into C code for transputers (small, inter-linked parallel processing nodes). Ironically this module taught me far more about neural networks than the neural network module I took. There are a number of Java packages which allow you to mimick the Occam constructs. Sadly you can't write Occam code and have it compiled down to Java - I used to love the simplicity of specifying SEQ or PAR in front of sections of code which were to be executed in SEQuence or PARallel. It was also the first language I encounted where white-space was important, just like Python. If you're feeling nostalgic you can still browse the language spec (pdf)

Diego on Symbian
Diego complains that Microsoft has a single SDK for all MS Smartphones whereas Symbian developers must get numerous different SDKs from Symbian and the device vendors such as Nokia, Siemens etc. "One download, one System" say Diego which kinda sounds like a modern day LOTR

I think there is also a number of other factors at work here although I do agree that developing for Symbian is still more focused on a particular device (say, 3650) than the whole family of devices, even with J2ME. But this is just a third-party developer's perspective. The other side of the coin is "how easy is it for the device vendor to port and customise the operating system?". Here Symbian seems to be winning outright both technically and politically.

If Nokia decides for technical and/or political reasons to go with Symbian there is already enough Nokia brand loyalty in the market to create an initial user base. And with an initial user base, developers will adopt the platform because quite frankly that's where the market is. The lack of a centralised developers resource is irrelevant to the users and us developers must grin-and-bear-it (or fix it?) so that we can access the growing Symbian market share. As a small mobile developer the (small) pain of development using multiple SDKs is offset by the market share you can access. Assuming my conscience could bear it, I still wouldn't develop for MS Smartphone because the market is in Symbian and J2ME applications in general [of course there is a huge market in Pocket PC apps]

Third-party developers are important but I don't believe that they have as much control in the mobile market as they do in the PC world. Users and vendors are far more important. Conclusion: Diego makes a valid point but I think there are wider and more important issues to consider.

oldie but goodie
You really can't beat a bit of early Skid Row - after catching a snippet of "18 and Life" on MTV this morning I went out and replaced my old (mildly shagged) cassettes with the budget CDs of the "Skid Row" and "Slave to the Grind" albums. And jeez there a new album coming out. 18 and life to go...

MIDP access to personal information
The Mobile Information Device Profile (MIDP) is the version of Java you will find on most new cellphones. Unfortunately it is impossible for developers to actually access the information part of the mobile device because the API doesn't allow you. So you think, "they must be working on fixing this blinding omission" and you trundle off to the Java Community Process site. Of the 51 specification requests (i.e. planned enhancements to J2ME) exactly none will enable you to check the users calendar, update their address book or in any way access the useful information. Instead we will be able to access and coordinate web services, listen to audio, watch 3D animations and create fancy graphs.

As useful and interesting as these might be, surely more interesting applications could be built if we were able to access the users information? We could schedule meetings between groups and check the dates with the calendar, store instant message ids alongside the contact information in the address book, and so on. In fact when you think about it, there isn't even a strong need for write access to these databases if security / integrity are of concern; just store the additional data in the J2ME record store and pull-in data from the calendar/address book as needed.

Am I the only one that has thought of this???

An aluminium keyboard. Why?
I always understood that aluminium computer cases were popular for two reasons: they're light to carry but strong and they help to dissipate heat. So why i God's name do you need an aluminium keyboard? Anyway, the reviewers seem to like it (too much money not enough sense?).

Diego has something interesting to say about Symbian and J2ME. Unfortunately, he's a bit busy right now.

stereotypes of the day
Nothing says "white trash" quite like a shellsuit and large gold earrings (except perhaps transparent white trousers and black thong on someone who shouldn't be wearing either - yuk). In much the same league (although less frequently seen) is the bloke with beer belly, braces and blue jeans - a.k.a German tourist (or a good impersonation of one).

That's just irrelevant stuff I saw today in Dublin.

My blogger code
B9 d t k s u f i+ o x+ e- l c--
Oh, this is far too geeky for a rural dial-up connection at the weekend.

Good to hear, but hopefully more to come:
REBOL/View has been open sourced. As Henri notes, this is the GUI element to the core REBOL language (not surprisingly called REBOL/Core). The killer feature about REBOL has to be the ability to email a webpage to someone with one line of code. From memory it was something like: send someone@somewhere.com read http://ideasasylum.com. Cool!

Hopefully the language will gain momentum if it was all open sourced - Carl Sassenrath (the architect) says REBOL/View an experiment so I'm hoping it goes well for them and we see REBOL gaining greater prominance.

I'd like to introduce
the trip blog which will chart our progress from here to... here, but via Australia and New Zealand. I've put the first post up there now but it'll only get really exciting around September when we leave.

Surely, the dumbest idea?
Apparently, we are going to sweep the global warming under the carpet (literally) by burying the offending gases. Err... how about we just stop producing so much of them? Oh and the best part was this:

"CO2 might be injected into nearby oilfields where it could help raise the amount of oil pumped out. [snip]
And it costs more fuel to capture CO2 ... Rather than use one unit of coal you use 1.3 units for the same power. Everything gets worse."

Can anyone else say "positive feedback" - more oil produces more CO², more CO² increases the oil production, which... you get the idea. There are times when I consider what I do for a job and how it helps the environment. It doesn't and that bothers me.

As Hilary notes...
Good plan: Go to the Blogger meet at the Davenport with Hilary and Mike.
Another good plan: Stay drinking there for 6 hours catching up and chatting. Also playing with Karlin's Nokia 3650 - cool!
Bad plan: Come home and decide to chop 5 tomatoes, 1 red onion, a chilli, and some coriander with the sharpest knife in the house. On Sunday morning I was relieved to discover that I hadn't lost a digit but this appears to have been more through luck (and now dented nails) than skill or sobriety.

See ya all at the Irish Bloggers meet. I'll be there with Mike and Hilary

small changes afoot
Ok, if you're reading this you also viewing the new site. Looks different doesn't it? Ok, no it doesn't but there are a few new features: comments, searching, a "jamie's mood indicator" and inline archives.

Comments are powered by dotcomments - I figured that I'd install it for myself since I've installed it for everyone else, and others are already queuing at the door! The searching is done my a little php script I wrote to scan each archive file, extract the relevent posts and highlight the matched words - give it a go and discover all the old stuff I've written! The mood indicator is just a little script which allows me to tell the world how/why I'm pi**ed off with it. Oh, there's some minor restructuring which will hopefully mean the the blog posts load faster instead of waiting for the junk on the right.

Feedback on any aspect of the blog is appreciated, just email me add a comment!

one of the best google hacks I've seen in a while
(via my brother)
1) Go to www.Google.com
2) type in (but don't hit return): "weapons of mass destruction"
3) Hit the "I'm feeling lucky" button, instead of the normal "Google search" button;
4) read what appears to be a normal error message carefully.

Of course, this will stop working in a while when the Google rankings are adjusted.

using Firebird or Mozilla?
you've gotta get Googlebar. Another reason I don't need IE anymore

nasty eircom
congratulations to the Advertising Standards Authority of Ireland (ASAI) for standing up to Eircom over their misleading (a.k.a. blatantly lying) broadband adverts.

Now I'm not one to diss Java
but freeroller.net (and all the weblogs it hosts) is so frequently down it's just not funny. I wonder what kind of uptime JSP-based blogs get compared to other blogging solutions? The problem, as I see it, is that although I have absolutely no control over problems at Blogger they only effect my ability to post, not my readership.

p.s. ok, I realise I don't actually have any readership but the point is still valid. mmm... talking to myself again.