Mac@Work : The Ultimate Business Machine

The
Ultimate
Business Machine

Technology, business
and innovation.

And, not least, about
the Mac.

Weblog • Archive • Cutedge

by: Bernard Teo

Wed 01 Dec 2004

The Story of Audion

Category : Technology/audion.txt

This is a good story, relevant to all software developers, about Audion, the MP3 player on the Mac that almost became iTunes. I really love its tone - not bitter and very generous.

Posted at 6:09AM UTC | permalink

Java and Objective-C Development on Mac OS X

Category : Technology/JavaAndObjC.txt

Objective-C code may look daunting at first, but it's really mainly C code with extensions to the language to deal with objects and message passing.

Both Java and C are small, compact languages that can (each) be learnt in one weekend - that is, if you're determined and start on a Friday night. Good books to use include the "Java in a Nutshell" book from O'Reilly and the Kernighan and Ritchie book for C.

The "Java in a Nutshell" book contains a good introduction to object-oriented programming and the advantages of using this approach to control code complexity.

I've got a few books on Cocoa programming but I think the "Learning Cocoa" book from O'Reilly is actually good enough to get started (I have the original version that was much maligned but the second edition by James Duncan Davidson should probably be better). I found some of the explanations mystifying (in my original version of the book), at least initially, but it contained very good, useful, sample code that you can adapt quite easily to do your own stuff. And, after a few rounds, things do become clear.

The important thing is to learn by doing - to plunge in and set yourself something useful to do by modifying from the given examples. Then it rapidly becomes clear what it is that you don't yet know, and from filling in the pieces, you start to create a mental model of how Cocoa works. I think there's a certain point where everthing clicks and you start to see the patterns in the way Cocoa works and learn to flow with it. (That's why there is a book called "Design Patterns".)

Now, I think we (Hai Hwee - who wrote Luca - and me) took a longer (and much harder) route into Cocoa by starting with Java. Cocoa's native language is Objective-C, and, though it is possible to write Cocoa apps using Java, you always have to translate the Objective-C calls into their equivalent idiom in Java, and there are far fewer sources of information about writing Cocoa code using Java. (As Aaron Hillegass says, in his book "Cocoa Programming for Mac OS X" 2002 edition - "Using Java with Cocoa"? "Don't ...").

But we have a lot of stuff we've invested time in building (database and web server stuff that call Oracle and MySQL) and it was easier to extend the Java code to work with a Cocoa front-end than it was to rewrite all that stuff in Objective-C. And where was the database solution on Cocoa in Objective-C, then? Nothing, not a single thing at all. No wonder there was not much going on for business applications on Mac OS X.

But, now, there's the possibility of taking advantage of CoreData in Tiger. One of the problems we have with Luca (besides the fact, I think, that it requires a good inventory module to make the system really gell for an end-user), is that it's too big a download. We can't assume that a user has MySQL loaded on his Mac. That is the target database that Luca uses. So we include a MySQL download. Even if a user has MySQL, we often have problems making Luca work with that existing installation. So it'll be great if we can assume that there is a database that comes with every Mac. And target Luca (and any other business application that we hope to build) to work with that common platform. And our applications could then become smaller and be easier to download.

So that's why I think we could expect to see an explosion of business applications on the Mac. But it all depends on how good CoreData is.

I do wish Apple could stop re-inventing the wheel and re-use stuff that are already out there - like all that JDBC stuff that work with almost every relational database solution on earth. Or, choose MySQL to bundle in rather than SQLite. Or make Objective-C work a bit more like Java in terms of memory management (all that "release" and "retain" code give me a headache).

We've now got two great platforms to create applications with on Mac OS X - Java and Objective-C, and Cocoa. It'll be great to get the best of both worlds. But part of growing up is to accept life for what it is and just move along. So on we go.

Posted at 6:08AM UTC | permalink

Mon 29 Nov 2004

May a Thousand Applications Bloom

Category : Technology/thousandApps.txt

"Deeds, like grains of sand spilt into the sea, disappear, leaving only an empty hand."

- Han Suyin

It's really quick, how time flies. Especially when you're doing something engrossing.

I've finished the Objective-C version of Postfix Enabler and it works exactly the same way as the current AppleScript Studio version, which shouldn't be surprising because they're both calling the same Cocoa framework underneath.

But what was surprising to me was that it didn't take that much more code to do it - almost the same number of lines of code in Objective-C as compared to the AppleScript code, or maybe just a bit more.

And I've reached this conclusion - if you're building applications in Real Basic or 4th Dimension, you really ought to take a look at building them directly in Cocoa using Objective-C.

You'll be surprised at how much faster and more productive you can be. You'll be coding closer to the metal, so to speak, and so your applications will run faster. But it's not necessarily harder - in fact, the opposite may be true. I can't praise the Cocoa framework enough - or rather, the thought that went into its design. So many things would have made their way in there simply because a programmer would have needed it - to work faster.

And I believe now that there's a way to get going on Cocoa quickly - where you can do a lot of productive things right from the start, and where you won't be daunted by how thick the Cocoa API reference guides are. In fact, if you find a way to understand the key concepts and learn how to look for what you want, you'll start to feel reassured whenever you go through a long list of API's - because that will mean that you're that much more likely to find just the right piece of code to help you do what you want to do.

And that has been my experience so far. I've been able to find almost all the things I need - or at least enough to think that I'd be able build quite a wide range of stuff with it.

Computers become useful because of the applications that run on them. I go the retail stores that sell Macs here and they don't give enough reasons for anyone to buy a Mac - they're always opening and closing windows, showing the genie effect, and showing iTunes, iMovies, and all that old stuff. They're preaching to the converted (or to the ignorant).

But we should use the Mac simply because there are so many productive things we can do with it - things limited only by our imagination.

So, may a thousand Mac applications bloom, with Objective-C and Cocoa.

Posted at 11:07AM UTC | permalink

Sat 20 Nov 2004

Postfix Enabler in Objective-C

Category : Technology/PFEinObjC.txt

I've been building a version of Postfix Enabler on Objective-C. What started out as a peek into Tiger to see how quickly I can get Postfix Enabler to run on it got diverted into a challenge - to see how quickly I can take the Postfix Enabler nib (Interface Builder) file and wrap Objective-C code around it, rather than the AppleScript I was using.

How did I get to Objective-C? Must be the effect of all those WWDC DVD's I sometimes run in the background while working, especially those about Core Data. If it works the way it's supposed to, and Apple continues to build on it, it's going to be something we cannot afford to ignore. It'll make the Mac an even better platform to build business applications on. And, if it doesn't? Well, we still have Java.

I've developed an AppleScript Studio course around Postfix Enabler because, if you take a look at it again, you'll find it covers most of the interface elements you're likely to want on a Mac-based application.

There are drawers, sheets, pop-up menus, check boxes, radio buttons, alerts, tables, data sources, progress indicators, password fields, interfacing with Unix, an ability to launch web pages, and checking if you're an admin user. Once you know these, you can build any application. (OK, maybe not any - I don't do graphics or music or 3-D).

So, by the end of this exercise, I will have an Objective-C application that covers most of the same interfaces, and more (e.g., I've always wanted to be able to read the output of the Unix job asynchronusly - so that I can show the user what's going on, as it happens, and not only at the end of the job, as in the current AppleScript version of Postfix Enabler. With NSTask and NSPipe, I've finally been able to achieve that). This will be just right to be used as the basis of a "Cocoa Using Objective-C" course. Objective-C is really so fun.

Posted at 5:34AM UTC | permalink

The Bull and The Vegetarian

Category : Commentary/BullAndTheVegetarian.txt

"Expecting the world to treat you fairly because you are a good person is like expecting the bull not to charge you because you are a vegetarian."

... Rabbi Harold Kushner

I remembered a quote like that came across my mail box some time ago. Fortunately, I've kept most of my Postfix and Sendmail Enabler-related correspondence. I found it in a note from Terry Bain. True enough. But is there a place in heaven for vegetarians?

Speaking of heaven, it's different hearing Eric Clapton's "Tears in Heaven" when you're a father of a boy yourself. It's easy enough to feel that stab of pain in - "Would you hold my hand, if I saw you in heaven?".

Posted at 3:34AM UTC | permalink

Wed 10 Nov 2004

The Job of Art is to Chase Ugliness Away

Category : Commentary/JobOfArt.txt

"The job of art is to chase ugliness away", I thought I heard Bono say in the U2 iPod special event. I had to rewind to be sure. Yes, it's there and I think it's a wonderful statement.

I was watching the event on QuickTime, while I was coming to the end of the book I'm reading by Jeremy Rifkin, "The Age of Access", and I think it's one of those days when everything resonates.

I was writing the other day about the idea of key concepts - about how it's important to learn how to learn quickly from scraps of knowledge.

But when you've built lots of these key concepts, you need an overarching framework to tie everything together, so you get a holistic feel for how everything is going, and thereby make your own sense of what's happening to our world today.

Jeremy Rifkin's "Age of Access" is one of those rare books that come along that does a credible job of tying lots of diverse impressions into a central theme, where you're able to see the patterns and interconnections.

We're moving into a time where we value the ability to access an experience more than the possessing of physical goods. For example, when I hear Bob Dylan's "You're a big girl now", lots of memories come flooding back and that's feeling I live for, no matter if I hear it off a radio or an iPod. What's important is to have the rights to turn it on when I want to - I don't need or care to own the CD.

Another example - I like the ability to whisk my family off to a few places quickly over the weekend - but I hate the chore of owning and maintaining a car. I'd rather lease one and I can see a day coming when somebody will use the web to coordinate the logistics to make the system work better than it currently does.

And I've stopped buying books - I bought enough to fill two life-times and they are just so much baggage when we shift house or office. I never finished Rifkin's last book "The End of Work" and it's now just collecting dust. But I found "The Age of Access" in the library, and I thought the worse that can happen is that it'll just go back into the hole if it's no good.

So, I'm glad I read it. But I'm not going to attempt a synthesis; just wanted to make the recommendation. This, and other books like Leonard Shlain's "The Alphabet versus the Goddess", have shaped my thinking about why Apple, alone among the technology companies, have grasped the true signficance of what's changing in our world today, and it's the company that is doing the most interesting and relevant things against this backdrop.

In what is coined The Experience Economy, art, aesthetics and alignment with spiritual goals take center-stage. The job of art is to chase ugliness away. And we're willing to pay for it when it does.

Can anyone watching the Apple U2 Special Event imagine anyone other than Steve Jobs being there on stage with Bono and company - for example Bill Gates or Sim Wong Hoo? Banish the thought. That's why the iPod's success is more than just massive spending on advertising, as Sim Wong Hoo thinks and is now trying to emulate. This may be Creative's waterloo.

Posted at 9:12AM UTC | permalink

Tue 02 Nov 2004

UW/IMAP - Postfix Enabler 1.0.10 Released

Category : Technology/UWIMAP2004Released.txt

I decided to release the latest UW/IMAP POP3 and IMAP binaries that I had built inside Postfix Enabler 1.0.10. Nothing's changed inside 1.0.10 except for the new POP and IMAP binaries.

I had earlier wanted to collect a few new features to put together to make a new Postfix Enabler release but I realised I'll never find the time to do that over the next four weeks.

The new IMAP binary includes the oft-requested change made to where it stores its mail boxes (so that it doesn't clash with Mail.app). It works so well with Mail.app now. It's fair to say that this is the way it should have worked, from the start.

I'm sorry I took so long to fix it. That's why I decided to release it immediately. This makes working with IMAP a much better and smoother experience and I can now understand why people would keep coming back to bug me for it. I hope the users are going to be delighted.

Posted at 7:56AM UTC | permalink

DNS Enabler 1.0

Category : Technology/DNSEnabler1dot0Desc.txt

I've got a more "useful" version of DNS Enabler done. It's meant to be used by people who don't know and don't care about things like CNAMES, MX records, PTRs,and A-RECORDS. But they know enough about the domain name system to want to be able to reference machines by human-understandable names.

The people who're downloading DNS Enabler now want to be able to do things like this, for a private 10.0.1 network :

This is because, for one reason or another, they may not be able to hit their server using a domain name, from a machine on their internal network, even though people outside, on the public network, can see the server just fine. For example, you may be acquainted with this problem if you're using the earlier batch of Airport Base Stations.

Running a DNS service on their server for the benefit of the machines on the private network will help these internal machines to "see" the server.

But you might want to set up a domain name service for a private network for a whole host of other reasons. For example, for doing demos or for testing, where you have no connection to the Internet, but want to simulate your own private Internet.

Now, the new version of the Airport Admin Utility allows you to create a private network on any of the three known ones (10.0.1.x or 192.168.1.x or 172.16.1.x). So, I've also allowed the user to set things up on DNS Enabler according.

DNS Enabler 1.0 allows a user to set up more than one server. And they can be on any other address beside 201 (which I had hard-coded in the "old" DNS Enabler). And, it allows you to set up more than one domain name pointing to the same machine so you can set up and do virtual hosting in the web and mail servers.

I find the ability to change, temporarily, which machines the domain names cutedgesystems.com and roadstead.com are pointing to very useful when I'm doing testing on my iBook. E.g., now that I'm doing a series of tests on the latest UW/IMAP binaries that I have built, using my iBook as both server and client. I find it very useful to be able to tell Mail.app on my iBook that cutedgesystems.com is now also on my iBook because I can re-use the mail accounts that I've already set up on Mail.app.

DNS Enabler will create the correct DNS entries and files - as far as I can tell (though I'm still learning and may yet get things wrong). So, after building it, I realised that it can also be used to set up a fully-functional DNS server for the public network, including handling the setting up of the Classless Inter-Domain Routing file (how's that for a mouthful), like the setup shown below:

And, finally, I'm finding that it works great as a teaching tool. The interface will allow me to explain to someone how the domain name system works and why it was meant to work that way. Then I can go into the system files that DNS Enabler generates automatically and show how all these settings get translated into the proper A-RECORDS and CNAME records, or whatever, that the Domain Name system requires.

I've yet to write the documentation. But if anyone wants to try it out before I've got a proper download page set up, you can write to me.

Posted at 7:29AM UTC | permalink

Mon 01 Nov 2004

A UW/IMAP Update

Category : Technology/IMAP2004Update.txt

I've built a new 2004 version of the UW/IMAP (POP3 and IMAP) binaries. And I took the opportunity to correct a mistake I made with the mail sub-directory path that led to problems like these (as excerpted from this page on the Squirrel Mail Wiki) :

"(1) Do NOT use the raw uw imapd that is installed by Postfix Enabler. This sets the imapd root directory on your server to ~/Library/Mail/Mailboxes, which is a folder that has special meaning for Mail.app. If you ever run Mail.app (ie Apple's Mail program) on your server it will put it's own stuff in that directory, imapd and Mail.app will fight about how it should be organized, and your life will be miserable."

With this new version, these problems should go away.

I should plan on releasing a new version of Postfix Enabler (1.1.10) soon (?) but I've got a series of courses that we're set on doing in a couple of weeks, so there's going to be a lot of things to do.

Posted at 4:48PM UTC | permalink

Address Book Plug In Update

Category : Technology/addressBookPlugInUpdate.txt

I've updated my Singapore Map Plug-In for Address Book so that it will try searching by Postal Code first. If it can't find a Postal Code, then it will try Street Name and Number.

You can download it from here.

I realised later, from monitoring the referrers on my webserver log, that Takashi Yoshida had already built something like this, so I am probably reinventing the wheel (except maybe for the one-click installer that I'm fond of building).

But I had fun doing it. And I did learn something more about scripting the Address Book that, I am sure, will come in handy in a future project.

Also, there should be a way to make this a general purpose Map plug-in - like, making it easy for people to set it up so that it'll work for any country/ mapping-site combination. E.g., providing a way for the user to add a new country and then enter the URL pattern of the relevant mapping site. The problem is to work out a smart way of deducing where the postal codes, street names or street numbers are in the URL of any mapping site.

Posted at 12:13PM UTC | permalink

Fri 29 Oct 2004

The Developer's Mac

Category : Technology/devTalk.txt

"Time is a jet plane, it moves too fast". So sang Bob Dylan on Blood On The Tracks. And I've got blood splattered all over my brain reading Mark Bowden's un-put-downable "Black Hawk Down", which I've reluctantly set aside just to get some work done.

Last week, at this time, I was wrapping up my stuff after doing the presentation at Apple WWDR (Asia)'s Mac OS X Development Seminar.

Leon Chen's put the photos from the seminar up on his .Mac page. It was pretty warm where I was standing and I think I looked a wreck. But I enjoyed it all the same. To think that I was once so shy, I'd die every time I had to get up and say something in class. But I was forced to overcome it when I was made an instructor while in the Army, and that was the best thing I ever got out of doing National Service.

Just one thing I forgot to mention during the talk, and I'd like to address it here. I was showing a series of demos based on the hostel system that we did, below, and I covered the variety of technologies that we exploited:

And the thing that may have crossed a person's mind, who's currently doing the same on the PC, is "So what?". And I wished I had stopped a while to address that.

I think the point I'd like to make is that we're really in the business of helping people understand information. That was the point of the demo - that we could tell a story from the point of view of a person booking a hostel bed, and of the person whose job is to operate the hostel, and of the owner of the hostel. The system proceeds to provide information from all three angles, so that the work could be coordinated. And that's why we have a system - to help people work better.

Now, it's very difficult to go a person, whose head is stuck in DLL hell, and ask, "Say, do you know what business we're in? Like, you know, the way Revlon knows that it's not simply in the business of making cosmetics" What business are we really in? It's like having someone coming to you and ask, "Do you know the true meaning of Christmas?" You'll say, "Go away, I've got to get this machine fixed and get back to doing some coding".

Now, if you could provide your programmers with tools that will just snap on and work, like the Mac, then you could go, "Now, do I have your attention? What else do you really need to do? So can I ask now, what's really going on here? What story are you telling your users? Can they understand it? And why does it matter?"

The tools that we use do shape what we are, what our priorities are, and what we care for. If we want to work like a craftsman, where we're told people enjoy using our products, then we've got to choose our tools carefully, and make sure we understand what we're really doing. And why we're doing it.

We're really in the business of helping people understand information better, through cutting away at complexity, and making the difficult things understandable. But we can't do that if we're always quarelling with our tools, or spending most of our time feeding it, or worse, intentionally complexifying things in the hope that that will protect our "marketability". In our age of the Internet, the truth will come out sooner rather than later.

Posted at 11:11AM UTC | permalink

Tue 19 Oct 2004

A Little Learning

Category : Technology/aLittleLearning.txt

They say that a little learning is a dangerous thing. But there's so much information out on the Internet, and so many things to understand, that you've got to figure out how to learn quickly from scraps of knowledge.

Take DNS, for instance. I've just finished a version of DNS Enabler that I can use for the upcoming developers' seminar. It's "better" than the one I had left around for download (in the article "OS X, Broadband, and the Airport Base Station"), in the sense that you can set the location of the server (it was always set at 10.0.1.201 in the "old" DNS Enabler). And you can set the IP address range (it used to be only for 10.0.1 networks).

And you can set up more than one machine on that network to be addressable by domain names. (In the seminar, I'm using this so that I can let the audience access both a PC and a Mac by domain names, if they want to see how well the same pieces of code run across the two platforms.)

So, at the moment, I have a DNS Server that is safe to use in a private network. But by building this, I now know what I don't yet know, and where to look, if I want to go one step further and make it a fully functioning DNS server that can be exposed to the Internet.

I remember the first time I attempted to set up a Domain Name Service, using MacDNS, back in the days of OS 9. I repeatedly brought down the Internet connection of our customer, out where we were camped then.

I only made progress when I bought QuickDNS Pro from Men and Mice (but they're not looking much like a Mac-type company anymore). They had a nice little DNS Expert application, which can be used to troubleshoot DNS set-ups, and that helped me built knowledge about all these arcane stuff bit by bit - including something called "classless reverse domains".

I remember the technician from our ISP throwing out these superior terms, when I had to tell him I was using a Mac, and I was expected to bow down and grovel because the Mac wasn't expected to have "it". I later learnt (rather painfully because this was before the days of the fire hose we now know as Google) that he was only speaking Unix-geek and that there was an equivalent way of setting this up using the more Mac-like QuickDNS Pro. Once I did that, plus a few other new things I learnt to check, the router stopped crashing on me.

When OS X came out, QuickDNS Pro didn't work on it. But I had OS X Server, and I could click the DNS-related buttons in Server Admin and go to the command line to see what changed underneath. That was my first taste of BIND. When I realised that even a stock OS X client had a working BIND built-in, I switched and ditched Server. Hurray, but I'm no DNS guru (yet!), and probably never will. But the thing is to be able to do productive things with each little piece of knowledge, because there is way too much data out there.

We've got to learn how to learn from the scraps that we get. I have people writing in wanting step-by-step instructions to do this or that. Well, I have the DNS and BIND book next to me and it's 600 pages long. I've got no choice, I can't read more than 10 pages at a time. The more important thing to do, always, is to create a skeleton framework of key concepts - something that will make sense to yourself and, more importantly, that you can use to try to explain to another person. Once you've caught the key concepts (and one should train oneself to do this very quickly), you can use that framework to hang the little pieces of information that will come your way, just-in-time, so that you can get things done when you need them.

Don't bother to memorise things. It's not an effective way of learning. Take the trouble to understand concepts. The problem I had with the ISP's technician was that he learnt things by rote. If he had understood things in terms of concepts, he would be able to describe the real issue, and then it would have only taken me a bit more time to say, hey, I have an equivalent way of doing that on my Mac, and thus solve the problem sooner.

I know I've got to get down from my soapbox but this is one way I've found that is effective for living in a world of information overload. Understand, don't memorise - that is what we should teach our kids.

Posted at 1:10PM UTC | permalink