El Aura

Divide and Conquer (solving the WWDC attendees problem)

Sat, 27 Apr 2013 02:55:00 +0200

Apple might have created a new record with its annual developer conference, WWDC, selling out in seconds. The conference with about 5000 attendees, probably was oversubscribed by factor of at least two. Since real-life Apple engineers are providing all the content in person, Apple obviously would not like to hold multiple conferences of this kind, nor would it like to hold it anywhere not close to Cupertino. Simply using bigger rooms (aka a bigger venue in general) doesn’t solve the access to the engineer problem and would degrade the social aspect that is very important.

There is basically no solution to the access to engineers problem for the one-to-one interaction. There is also no solution to the vibe a one event per year with all the people in one venue together creates. But Apple’s problem here occurs in all fields were people come together to one physical place to exchange information, network and socialise. And they are all tackled by dividing things until they have reached a manageable size again. It doesn’t solve the access to engineer and the vibe problem but it is a tried and true technique.

Applied to the WWDC this means:

Create technical subdivisions and hold them at separate dates. Yes, this means for Apple different parts of your company are disrupted at different dates. But maybe that is less disruptive overall.
Hold conferences on different continents and countries at different dates. That is the more difficult one to implement as it would mean for the same engineers to attend multiple conferences (and to travel long distances). To make that workable, probably the non-US conferences would need to make do with much less engineers (or even via video conference).

In a sense, Apple is already did that with its Tech Talks. My suggestion essentially would mean be to give them a higher status to hopefully create events that have some of the vibe and networking of the WWDC. And this should apply to both the technical and geographical subdivisions of WWDC.

Of course, Apple would probably still want to have one major conference which would make it difficult to have human resources allocated to both the major and the multiple minor conferences but maybe a reasonable compromise could be found.

There is also the option to implement the geographical subdivision via life video streams but at only a very small number of venues (eg, one in Europe and one in Asia). One could have a few in-person events at the other venues which could be unique or life-streamed to the San Francisco venue. The real question here is whether such subdivisions, especially the geographical ones, would attract enough people and give them a good enough experience, ie, whether the lack of physical speakers (and the access to engineers labs) or the scaled down technical divisions conferences would still make this a worthwhile experience.

No, Apple has not changed (yet)

Wed, 30 May 2012 20:11:31 +0200

“Ever since Steve passed, the company has seemed to be off course – not financially – but in regards to vision. Even before Steve passed, we saw the Final Cut Pro debacle – turning one of the most successful professional video editing applications into a glorified version of iMovie.”

Scott Bourne

Apple’s mojo in the second Steve Jobs era was always to make some bold steps and then refine. Sometimes the refining got stuck and the product was either abandoned or another bold step was made. Final Cut Pro X fits 100% into this pattern. Calling it a glorified iMovie is a dismissive polemic (ie, lying by exaggerating for the personal satisfaction an attack on something yields). The current version of FCPX is much closer in features to the last pre-X version than to iMovie.

In regard to Aperture, nothing much is new. It is just the same trend that has been there all along with LR receiving more frequent major updates and having an edge regarding image adjustments and Aperture an edge regarding organisation. It is just that the gap that has been growing slowly but steadily (in regard to image quality but also to some extent speed and reliability) has become now big enough to being noticed more easily. Aperture had a head start but not one that was long enough to become really entrenched and its effect has all but disappeared.

Apple’s secrecy has also hardly changed over the last decade but when it brings great products and surprises, we like it and when there is a longer period without new products, we hate it.

All that does not mean that LR might now not be noticeably ahead of Aperture in a lot of areas. But we don’t need to spice up that conclusion with how things were better under Steve and anyway everything was better in the past. Because they generally weren’t.

P.S.: Raw support in Aperture for new cameras has noticeably speeded up this year and has beaten LR for a number of cameras if we don’t count pre-release versions.

The short story of Instacast vs. Downcast

Thu, 29 Mar 2012 21:11:00 +0200

I was reluctant to switch away from iTunes for my podcasts, partially because of a somewhat diffuse feeling that my computer is still my digital hub (not least because of oodles of storage that is backed up in half a dozen different locations) and partially for the simple question of how to manage a back catalogue (and backlog) of podcasts that simply don’t fit on my iPhone.

And with automated checking for new episodes, automatic deletion of listened to podcasts, and one-click WiFi syncing, letting iTunes handle them seemed like a simple enough solution. Except that WiFi syncing at work was far from one-click. Sometimes the iPhone would not manage to log into the primary WiFi network, meaning I had to switch both my computer and my iPhone to a secondary one and enable VPN on both. And starting the sync on the computer would basically never work, starting it on the phone would allow for the handshake but still needed to be followed up with a sync from the computer. On most days, I also had to restart my iPhone for even the handshake to occur. Overall, I probably spend five to ten minutes every day to do that WiFi sync at work.

Thus, I reached out to standalone iOS podcast clients, leaving only the highly backlogged podcasts and those with a large back catalogue to be managed by iTunes. Here is short list of deficiencies in Instacast that led me to choose Downcast:

No chronological listening to episodes
No way of seeing from a list of episodes which is the one currently playing
An extra tab is required after having selected an episode for it to start playing
General unreliability in deleting listened to episodes and downloading new episodes
No list of currently downloading (and queued) podcasts (at least I could not find it)

BTW, both apps failed to import existing subscriptions from an iTunes-generated .opml file, be it via e-mail or via Dropbox.

Save As vs. Duplicate in Lion

Thu, 01 Dec 2011 18:28:00 +0100

Having used Lion for more than a month now, I do not understand at all all the bluster about the auto-save behaviour and the disappearance of the ‘Save As’ command. Not having to think about saving just feels natural to me. And duplicate also feels like the more natural way of doing things, since we always used ‘Save As’ as a means to duplicate a document.

Of course, the lack of a default shortcut might trip people up but it is the price to pay for having a stretched-out transition from the old behaviour to the new. Apple probably should have used cmd-D or cmd-shift-D but some people, including John Gruber, feel more comfortable with keeping their old habits, i.e., cmd-shift-S, and/or want consistency among applications as a lot of ‘legacy’ applications still have, and will have for quite some time, a ‘Save As’ command with an associated cmd-shift-S shortcut. Maybe Apple felt no shortcut was better than picking one of the two (S or D), particularly since it is easy to add one yourself.

Installing Lion on a RAID volume

Tue, 09 Aug 2011 16:33:00 +0200

Installing OS X Lion automatically creates a new invisible recovery partition on the volume the target partition resides on. Unfortunately, if the target partition is a software RAID volume created by OS X’s Disk Utility (or probably any kind of RAID), the installer is unable to create this additional partition (or at least without destroying the software RAID) and might fail.

Reportedly, you can install on a software RAID if you can boot the computer which hosts the RAID from another volume either via the official installer or by cloning a Lion installation/partition (belonging to the same computer) from this separate volume onto the RAID volume. Unfortunately, for computers that ship with Lion, the App Store version of Lion does not install (they need a slightly different build). Thus, you need to first obtain a Lion installer for this particular Mac and ‘install’ it on the separate volume. Once that is done you can create the software RAID with it and either install with the previously obtained machine-specific installer or clone to the RAID.

Currently the only way to obtain this machine-specific installer is via Internet Recovery Mode (http://support.apple.com/kb/HT4718). You have two options to proceed:

——— A ———

boot into Internet Recovery Mode
start installation onto separate volume (this takes time because it first has to download the installer)
boot from the installation on the separate volume
create Soft-RAID
clone the installation you are booted from onto the Soft-RAID

——— B ———

boot into Internet Recovery Mode
create Soft-RAID and try to install onto it (this takes time because it first has to download the installer) and will ultimately fail with an error message
(while still being ‘booted’ from the Internet Recovery Mode) restore the machine-specific InstallESD.dmg (that the installation attempt dropped somewhere onto the Soft-RAID) onto the separate volume,
boot from there (ie, boot from the InstallESD.dmg which is the installer)
install Lion onto Soft-RAID

For some reason, this will not install Disk Utility onto the Soft-RAID (probably because the actual Disk Utility binary usually sits on the recovery partition which does not exist on the RAID.

————

Method (B) is taken from http://translate.google.ch/translate?hl=en&sl=auto&tl=en&u=http%3A%2F%2Fwww.skgm.de%2F%3Fp%3D405, method (A) is a patchwork gleaned from multiple sources including comments from the previous link and I thus might have jumped to conclusions with it.

My top XX iPhone apps are not your top XX apps

Fri, 24 Jun 2011 01:11:00 +0200

I am amazed when I hear that people only use a handful of apps. Loosely sorted by usage frequency, these are the third-party apps I use regularly:

BBC News (listening to the BBC Wordservice)
RSR (listening to my favorite music radio station)
Reeder (RSS client)
Shazam (identify music)
Skype (VoIP, lowers your phone bill)
Articles (Wikipedia client)
Simplenote (synchronised notes across all your devices)
SBB Mobile (public transport schedules)
ZVV Fahrplan (ditto, better at finding local bus stops)
Swiss Map (high quality topographical maps)
Dropbox (easier than emailing you files you might need on the go)
Weightbot
eBay (watching auctions on the go)
Amazon.de (looking up book titles, authors and lots of other stuff)
The Economist (reading or listening to your favourite newspaper without having to carry anything extra)
DB Navigator (public transport schedules for another country)
LEO (the best online dictionary between German and six other languages)
Ninjawords (the best English dictionary)
PCalc (clean and useful RPN-capable calculator)
i48 (when you need the full power of an HP48 calculator)
Flashlight-4 (use the LED flash as a flashlight)
IMDB (query the IMDB movie database)
Kindle (e-book reader)
Telefonbuch (German telephone directory)
mobility car (carsharing access app)
Public Toilets (find them wherever you are)
iWeather.ch (rain radar for Switzerland)
Star Walk (find any star or constellation, or just the moon or the sun ;))
Convertbot (convert between units)
Currencies (convert between currencies)

Apple Discussions - making sense of the nonsense

Tue, 19 Apr 2011 17:06:00 +0200

Apple’s new discussion pages offer a lot of options viewing the content. But as usual, just adding options rarely makes for a useful product, something that Apple is normally keenly aware of. Alas, this time they threw it all out and did not even attempt to provide a useful default view. Here is my overview of the available options, bold text referring to actually useful elements:

A) Content

Apple Featured Topics (a global selection of threads Apple considers useful, unlikely to be useful as way too much and way too widely spread)
~~Community Annoucements (this category seems to have disappeared again, though I still have it on my homepage)~~
Recent Activity (current global background noise)
Recent Announcement Posts (announcements by Apple)
Recent Content (the same as recent activity, maybe minus announcements?)
Recent Discussions (the same as recent content, minus the basically non-existent documents/user tips)
Recent Documents (currently only two single entries from this month)
Top Liked Content (most often user-voted posts)
Top Rated Content (the same as top liked content)
Unanswered Questions (current global background noise minus the small proportion of answered questions)
View Document (sole way of editing any documents, if you did not find place for this widget among all the other 50 widgets on your home page, you cannot edit documents, not that you could create documents, that seems to be broken at the moment)
Watch A Tag (put a subforum on your homepage, would work if people actually used tags, unlikely to happen)
Watch A User (unless you stumble upon a user that mainly posts stuff useful to you, which is pretty unlikely, just some only slightly filtered noise)
Your Content (only way to get to all your posts, or rather all threads with posts of yours, hopefully all, only useful when put at the lowest position of your homepage as it will be very, very long since it has no page breaks)

B) Places

Communities (put the list of forums on your homepage)
Places (same as communities but puts forums you have joined on top)

C) People

Newest Members (if need some inspiration on choosing a username, a random list of new names)
Recent Activity (same as recent activity, very original)
Top Participants in Communities (across all forums, just look at the points and you don’t need this list)
Watch A User (same a watch a user)

D) Other

Formatted Text (add random text to your homepage, eg, ‘This list below is completely useless)
HTML (add random anything to your homepage, eg, if you prefer AD to Facebook)
Recent Bookmarks (delicious for Apple Discussions)
RSS Subscriptions (if none of the above gave you the right mix of AD content, add RSS subscriptions to it, if you don’t have an RSS reader, you use your homepage here as one)
Your Bookmarks

————————————

So, besides forcing users to wade through 24 categories to find the handfull of them that are actually useful and then to configure the categories further it offers a parallel universe with the controls in the top-right menu:

1) Your Stuff

Bookmarks (ditto as above)
Discussions (similar to your content, except incomplete)
Documents (presumably the same as edit documents, except that documents are non-functional at the moment)
Communities (same as places)
Notifications (that is new or is just another name for announcements?)
Preferences (yeah, that is new, but creating your homepage is not part of your preferences)
Profile (another section of preferences)

2) History

(replacement for your browser history limited to AD, needed as a lot of pages do not get their own URL, partially overlaps with your content)

3) Browse

Announcement Posts (ditto as above)
Bookmarks (ditto as above)
Discussions (ditto as recent discussions, ie, useless)
Documents (ditto as above, and useless as non-functional)
Communities (ditto as above)
People (ditto as above)
Tags (ditto as above)

————————————

And to top it all, people who are just browsing (ie, not logged in) get only a search box and a listing of the forums and the posts therein. They cannot see announcements, bookmarks by others, top-rated posts, featured posts, or documents. Great job.

Is Apple's grab for a cut of iOS revenue worth it?

Fri, 18 Feb 2011 04:18:12 +0100

Three questions come up:
(a) Are we fine with Apple driving off all competitors for any market for iOS apps it chooses to compete in (iBooks does not have to pay a 30% tax to Apple, Amazon however has to, leaving very little room for any iBooks competitors, and that is before counting the private APIs and customer information advantage iBooks has)?
(b) Do we think it is both fair and good for the platform we love so much that any iOS business can be destroyed or forced to withdraw from the platform by Apple at any moment?
(c) Are iPhones and iOS worth it to pay 43% more for every single service or product used on it, including products like Skype calls (that is if companies find ways to de-bundle their offers in a way that allows them to include Apple’s 30% into their prices, if they do not, they will either quit or raise prices for everybody)? And are iPhones and iOS worth this potential debundling?

Apple's deal with the Appstore

Wed, 19 Jan 2011 14:45:00 +0100

With the original Appstore for the iPhone, Apple offered us a deal: We are only allowed to install applications through Apple’s vetted Appstore but if we would not charge our users, Apple would not charge us for hosting the apps. At least as long as we respected a set of overall reasonable rules. One of the least reasonable rule in many eyes was that apps should not compete with the core functionality of the iPhone. But even here, public pressure and inquiries by competition authorities finally led to the acceptance of apps like Google Voice.

With its, as yet announced and unreleased, new subscription service for periodical publications, Apple is now expanding what it sees as core functionality and thus starting to put that original deal into question.

Apple's attempts to monetize on free apps

Tue, 18 Jan 2011 15:29:00 +0100

The recent private message sent by Apple to several European newspapers that it wants all access to paywalled content channelled through its own upcoming subscription service instead of going through free apps and an access code sold via the newspapers websites might primarily be motivated by giving its planned Appstore subscription service as much scope as possible. But it also serves as a first step in ending the concept of free apps.
While free apps are likely to remain available for the foreseeable future this marks the first step by Apple to start cashing in on the profits generated by free apps. While selling a service by channels other than the Appstore and then allowing users to access this service via free iOS apps might seem as an illegitimate way of bypassing Apple’s store and thus undermining its financial foundation, attempts by Apple to prevent this will come with a number of serious challenges.

For once, the notion that a company has to pay to the OS or device manufacturer to run an application on them seems out of touch with free and highly competitive markets other infrastructure devices like desktop computers, telephones, or transportation sectors have seen. Only in areas with little or no competition, eg, mobile devices with closed operation systems or cable television, do companies have to pay custom prices just to be present. No company has to pay a special price to be reachable by phone, no company is paying Microsoft to be able to install applications on its OS, at least not beyond standardized, non-discriminatory prices.
Determining what price a company should pay to be present on an iOS device, moreover, is a tricky business. Apple’s standard rate of 30% of income generated from selling an application seems like a simple enough formula. It however becomes close to impossible to apply when services are bundled, eg, combining online access to a newspaper’s (paywalled) website and to the same content via an iOS app. Of course, unbundling solves that issue but it mostly is not something that is attractive for customers, paying once to read a Kindle book on an actual Kindle and then pay again to read it on an iOS device.
In other areas it might be easier to determine the revenue associated with an iOS app (and apply Apple’s 30% take on it), eg, advertising inside an app. But even here, any bundling (eg, selling ads inside an iOS and Android app as a package) would conflict with it. Where it truly becomes impossible to gauge on what to apply a 30% cut on is for apps which serve as advertising or promotion themselves, eg, an airline’s app.
In the short term, Apple’s change of tack might only result in a set of forced unbundling, ie, selling newspaper iPad access separately from its online access. But that will only work for a limited number of services and it would be a very messy affair if applied on any larger scale (bidding through Ebay’s iOS app would direct some kickback fees to Apple but just watching auctions would not?).
Apart from some very limited and unpopular unbundling or likely illegal attempts to force all advertising in iOS apps to go through them, Apple will have little success in getting its share of the profits from free iOS apps. The company would be much better off to simply charge for the hosting and vetting of free applications (while exempting not-for-profit organisations and low volume apps). Applying additional fees to successful companies will be a very complicated affair and would likely be counterproductive.
Apart from the hosting and vetting costs, Apple should thus rather refrain from further attempts to monetize the Appstore as the availability of apps is something that in itself provides huge promotional support to its iOS devices.

Eye-Fi: How not to write software

Wed, 25 Aug 2010 20:44:00 +0200

The two main bugs I ran into were:

(A) When connecting to my Eye-Fi card for the first time, the Eye-Fi Center and the Eye-Fi webpage refused to let me create a new account, instead asking me to log into my account using my not-yet-existing credentials. Support told me that this should only happen with an already registered card but the Eye-Fi Center clearly showed my card as unregistered (see screenshot post below).

How did I solve it? No idea, at one point clicking somewhere in the window changed the screen from login to register. Trying to repeat this, it went to register right-away. In the end, it might have been that the ‘register’ screen only came up when using the version of the Eye-Fi Center software supplied on the card and not a downloaded, updated version (which the software initially prompted me to do before I even tried to register). Or it was just luck that it worked at some point eventually.

(B) Trying to set up a card became almost impossible on the second day (again it might be that on the first day I used the older card-supplied version of the software but I think I was working fine with the up-to-date software on day one as well). Basically, whenever an Eye-Fi card was directly connected to the computer, logging into my account (to start doing any changes) worked only for one or two seconds and then I would be kicked out again with the login screen appearing as an overlay. Eventually I managed to change my settings, but it was a real pain: hit login button, click field to edit, type one or two characters, be kicked out again, repeat until all changes are done. Every time, I was kicked out, the focus was lost and I had to re-center it to the field in question. A few times, I was able to bring up the settings window (via Eye-Fi Center -> Settings -> Eye-Fi Card) on top of the login screen and it stayed there.

Other minor annoyances are:

(1) Installing the Eye-Fi software required quite a number of Adobe Air updates (which required a felt ten mouse-clicks in total).

(1) Starting the Eye-Fi Helper requires typing four characters into the Spotlight filed (+ four arrow downs) or alternatively tying eight characters.

(2) After starting the Helper, its menu bar icon gives me the beachball for five to ten seconds.

(3) Almost every time I start the Eye-Fi Helper I need to re-type my credentials

(4) For some period, typing the password in the login screen was a real pain as the focus switched from the password field to the user name field after every single keystroke.

(5) At some point the helper software kept crashing on start-up, support told me to re-install, that fixed it for the moment but had occasional crashes since then.

I would not be surprised if part of the blame actually lies with Adobe Air, but overall this is one of the worst pieces of software I have used.

Eye-Fi software presenting me with an overlayed login window...

Wed, 25 Aug 2010 20:34:01 +0200

Eye-Fi software presenting me with an overlayed login window even though I am already logged in (discernable from the logout menu option and the half-visible settings window in the back).

Eye-Fi software refusing to offer me a registration option on an...

Wed, 25 Aug 2010 20:32:19 +0200

Eye-Fi software refusing to offer me a registration option on an un-registered card.

Eye-Fi and iPad

Wed, 25 Aug 2010 19:55:50 +0200

With the application ShutterSnitch it is possible to transfer images directly from a camera wirelessly using an Eye-Fi card onto an iPad. Using a mobile wireless router (MiFi, Huawei E5830) this also can be done in the field, even with no internet access. But setting things up is not exactly obvious. Once the basic Eye-Fi settings have been defined (see other post), here is how to switch between uploading to the iPad and the computer:

To switch to uploading to computer:
(1) Connect computer to the internet
(2) start Eye-Fi Helper, wait ten seconds
(3) start Eye-Fi Center
(4) enter credentials and login (they might be already filled in or they might not)
(5) click on settings cogwheel under Devices on the left
(6) switch to Photos, click on ‘Enable Uploading Photos to computer’, confirm with Save, if already checked, uncheck and re-check
(7) confirm switching from other device to computer
(8) quit Eye-Fi Center

To switch to uploading to ShutterSnitch on iPad:
(1) Disconnect computer from network or shutdown Eye-Fi Helper
(2) Connect iPad to the Internet
(3) Open ShutterSnitch and under Options: Set up Eye-Fi access and follow instructions (five taps in total)
(4) Restart ShutterSnitch (not always necessary)
(5) Switch off original, internet-connected WiFi network
(6) Exit ShutterSnitch and switch to local WiFi network
(7) Repeat the Eye-Fi ‘pairing’
(8) Enter a collection and start shooting

(Steps 5 through 7 are only necessary if a local WiFi network without internet connection is be used.)

Getting files on and off the iPad

Tue, 01 Jun 2010 00:26:41 +0200

Von meinem jetzigen Wissenstand her, würde ich sagen, die einfachste Methode Sachen auf den iPad zu bekommen ist Dropbox (sehr schnell wenn Computer und iPad im selben Netzwerk sind, vollkommen automatisch, funktioniert wie ein einfacher Ordner auf dem Computer).
Auf dem iPad aus der Dropbox App heraus öffnen, und dann auf iWork.com exportieren (dies geht über das Internet, kann daher langsamer sein). Auf dem Mac aus dem iWork.com Browserfenster wieder auf den Mac verschieben.

Der Hinweg über die Dropbox ist wie als ob man von beiden Geräten auf den selben Folder zugreift (ein Ordner der immer automatisch in sync ist auf allen Geräten). Leider kann Numbers auf dem iPad nicht zurück in die Dropbox speichern, deshalb der Umweg über iWork.com. Und leider kann man iWork.com nicht also lokales Filesystem (wie Dropbox) mounten, daher der Umweg über den Browser.

Im Moment schickt man also ein File im Kreis herum, aber der Aufwand hält sich durchaus in Grenzen. Die größte Einschränkung ist das man sich auf Fileformate beschränken muß, die iWork lesen kann und daß man immer in ein iWork file format abspeichern muß (oder ein PDF).

Es gibt natürlich noch die Möglichkeit per E-mail Files zu verschicken, die auch in allen anderen Apps (z.B. Penultimate, Omnigraffle) funktioniert, aber natürlich auf Dauer etwas unhandlich ist. Und iTunes natürlich, aber da muß man immer von Hand einstöpseln und Files ‘freigeben’ wenn man ein File verschieben will.

Mercurial's real main advantage: merging based on changesets?

Sun, 28 Mar 2010 00:19:22 +0100

With a decent experience in Subversion (albeit in tiny teams), I read this tutorial on Mercurial with great interest. My general take is that the real advantage of Mercurial is its merging tool, ie, one that actually works and the fact that it works is because it is based on changesets not on final states. Plus the ‘UI’ of doing multi-level repositories and merges looks more userfriendly and straightforward than in Subversion (where this multi-level approach can be achieved with personalized branches or other third-party tools).

Still, I feel support for Mercurial in terms of clients and hosting services (within and outside institutions) is not on the same level as for Subversion yet, though it is coming (eg, http://www.shelfcloud.com/blog.html). But the client problem might be fatally linked to the GPL-nature of Mercurial. I have also not seen yet how to best split up and recombine large projects (which in Subversion can be ‘done’ by only checking out parts of a repository).

Nikon Tele Lens Prices

Thu, 17 Dec 2009 00:00:00 +0100

Nikon Tele Lens Prices:

(Click title for graph)

Very crude guide at what Nikon tele lenses cost, based solely on the theoretical area of the front element needed for the noted aperture at maximum focal length. Line is a third-order polynomial fit (parameters in box), prices based on early December 2009 B&H website. Red numbers are deviation of actual price from theoretical one.

As it can be seen, zooms with current technologies (VR + AF-S) are above the trend, as one would expect, ie, zooms being more expensive than primes. Non-VR lenses like the 300 mm f/4 AF-S is below the trend.

Red line indicates rumoured 100-500 mm f/5.6 lens. Anybody venture a guess for its price?

Creator codes

Wed, 07 Oct 2009 23:22:00 +0200

John Gruber makes all the right points (http://daringfireball.net/2009/10/congrtlns-osx) but fails to mention one crucial one: There never was an (official) interface to set creator codes. He correctly concludes that:

“The creator code long ago stopped meaning “the app that created the file”, and instead meant “the app this file should open with by default”.

If people had used the creator code to mark individual files as destined for one application and if this ability were now gone, there would be reason to complain. But no application I have ever seen had a check box: ‘Set creator code’ and no part of the Finder allowed you ever to set the creator code. Only third-party utilities did let you do this and and required that you knew the precise set of characters. Hardly different from how the situation is now.

To follow Gruber’s own reasoning, if the preferred application for any given file nowadays is more likely than not to be an application other than the one that created the file, surely, creator codes are more likely to result in an unintended behaviour than not. Particularly, given that the user has no choice whether an application sets the creator code or not except for a file-by-file manual override later. There are good reasons why basically no application still sets the creator code (if you subtract those using a custom file name extension, making the creator code already pointless).

Boost, Bjam, Xcode for Dummies

Wed, 01 Jul 2009 16:37:00 +0200

Boost (boost.org) is an extensive set of open-source, peer-reviewed C++ libraries. Associated with it is a compilation/make tool called bjam. Here are some quick and dirty instructions to get it running:

Installing bjam:
- Install bjam executable in your path (eg, /usr/local/bin), get it from: http://www.boost.org/doc/libs/1_36_0/more/getting_started/unix-variants.html#get-bjam
- Install boost-build into /usr/local/share (no compilation necessary)
- add to /usr/share/boost-build/user-config.jam one single line (OS X specific, other *nix gcc): using darwin ;

Installing the boost libraries:
- for header only libraries, copy them anywhere (common location is /usr/local)
- for compiled libraries, install with:
> sudo bjam install —prefix=/usr/local/ —toolset=darwin
(for other systems choose the appropriate toolset, eg, gcc)

Configuration and invocation:
- compile with bjam (or bjam release, ie optimize the code)
- you need a file called ‘Jamroot’ in the root of your code directory, it minimally contains:

exe nameofexecutable : sourcecode.cpp ;

- to use the boost libraries you have to add after another ‘: ’ but before the final ‘;’ an <include>/usr/local/include/boost-1_39 (ie, containing the path to the boost libraries, maybe repeat with /usr/local/lib/boost-1_39)

In Xcode:

- in the project window, click on the project name, get info, add to ‘Header Search Paths’ the path to the boost folder, eg, /usr/local/include/boost-1_39, uncheck the recursive box
- select the executeable, get info, add to ‘Arguments’ str

Refinements:

Add an environment variable to your .bash_rc file (or similar) (eg, add the line: BOOSTROOT=”PathToBoost”) and use $BOOSTROOT in the Jamroot file and set up a Source Tree in Xcode’s preferences with that name and value.

References:

http://www.boost.org/doc/tools/build/doc/userman.pdf

http://old.alanhogan.com/code/boost-xcode

http://lists.apple.com/archives/Xcode-users/2008/Jul/msg00068.html

http://stackoverflow.com/questions/749027/how-to-add-a-global-include-path-for-xcode

http://markmail.org/message/mnpajba5undaq5xi#query:bjam%20%22include%20path%22%20%22get%20started%22+page:1+mid:ow2vqhsada5zgudx+state:results

Realplayer and Time Capsule

Fri, 13 Feb 2009 00:44:00 +0100

A lot of people are getting socket errors on all Real streams when connected via a Time Capsule. As always when certain internet connections systematically do not work while most others do, suspect a port issue in your internet connection pipeline. At one point, I thought only Realplayer 11 was affected but other users report problems with version 10 as well. What helped for me was expanding the UDP port range (in the Transport section of Realplayer’s preferences) from its default 7070-7071 to a very wide 6970-32000 AND unchecking the multicast option for RTSP connections (manually configuring the RTSP settings). Other suggestions of switching to the http protocol only did not work (possibly the server side has to support this different protocol, so this might work with some servers), forwording (portmapping) UDP 7070 in the TC also did not work. Apparently, the TC blocks some (incoming) UDP traffic but this is far from transparent. Well done Apple or Real, take your pick.

Even more severe problems can occur when connected to the internet via a 3G modem, as most (all?) mobile service providers (MSP) only keep a very, very limited number of ports open (VPN was very shaky, Matlab license server did not work, sending e-mail was touch and go).