Why WWDC will never be like WWDC again...

WWDC stands for World Wide Developer Conference. An event annually held by Apple in San Francisco. As the name implies, it supposed to be "world wide". However due to the popularity of iOS, the event is being flooded with developers and since 2 years was being sold out. In 2010 however things changed.

Europe received invitations for WWDC at 18:05. By the next morning, WWDC was sold out. In other words, any serious developer who has to organize hotel and long distance international flights to San Francisco has absolutely no chance to get to WWDC. He might not even have seen the invitation before it was sold out because of the timing it was sent. The result of this will be that WWDC will no longer be World Wide anymore. It will be 99.999% US only while more developers who develop for iOS are abroad, while more revenue is being produced with Apple devices outside of USA, I consider this quite a drastic shift.

Unparalelled Support.

Addendum 2:

Even though most issues had been solved. Our virtual mailservers running CentOS 5.5 and Kerio Mailserver and nothing else still dissapear from time to time. You can't ping the virtual machine, you can't run the Parallels Management Console (Can't connect) and only a reboot of the host system (a XServe running MacOS X 10.6.4) fixes it. The system log doesn't show anything obvious. *sight*.

Addendum:

Parallel's has addressed the major issues in Parallel's Server for MacOS X Version 4. There are only a few graphical GUI glitches left which are by far not critical. So if you run version 3 and have issues. Upgrade to 4. It will make your life a lot easier.

The Linux variant runs like a charm on our IBM blade server but still has issues on the other machine I used. I've given up on that one and write it off under "bad motherboard".

---------------

As we are building a large virtualized datacenter, we are testing virtualisation solutions. As we want to offer MacOS X Server hosting and want to do "real stuff" we looked closer into Parallels Server for Mac OS X. About 2 years ago I tried this before and it failed. It crashed, freezed etc. But as time has passed and Parallels Desktop usually is working fine, I thought I give it another try. But big problems arised.

1. You configure 8 virtual machines on a XServe with 20GB of ram and big fat harddisks. All looks great until you reboot the machine. The virtual machines won't start. So you check the box to autostart and save. But oh'oh, it is not starting neither. The settings don't get stored. You get prompted with weird error messages sometimes pointing to some "credentials issue" but if you simply do it again it doesn't happen. And some virtual machines look like "stating" but they arent. So you can't stop or start them. Rebooting the server is the only option.

2. Ok so let's try it on a Linux machine with the Parallels Bare metal server. Sounds like "kernel and app and nothing else" should be slim and working. After doing this on a big fat PC with many disks it all looks great until you reboot the machine a few times and your boot disk gets totally corrupted. Ok sounds like a SATA disk issue, new controller, different disk, same problem. But now the box is in a datacenter 3000km away from home.

3. Contact support. Well go to the webpage and try to contact them. Contact us requires registration. Well I am registered but I can't login. Password recovery works but logging in is still bringing you back to the login screen. What now? theres no way to contact support.

Doing the same with Firefox instead of Safari reveals that they have a buggy web designer. Well, ok now you are there. You can do a life chat. But oh, theres no product Parallels server there. So I have to choose Desktop. And then you are #1 on the list for the next 30 minutes. My patience is not so long...

4. Next thing the app crashed once so it created an automated submission and now I have a ticket on their RT system. Great. Finally I can explain my issues to someone. Its a priority #1 case as everything stops after a reboot so its a mission critical issue. You get an answer saying you get a reply within 24h. A day later nothing happened. After writing 5 additional add-on's to the support case, I get some lousy answers which gives a script (which isn't attached of course) to start the servers to be installed as a LaunchDaemon. While this is a ugly workaround at least for the not starting stuff, its no option for the hanging processes at start.

5. there's a nice checkbox in the config tool saying "iPhone support". if you check it, it says its not implemented. WTF. Why have a checkbox at all if its not supported??

6. You can't install CentOS 5.5 on the virtual machine due to some graphics driver issue. As this is a server, there's no graphics needed. plain old text mode however doesn't work neither. But why does it work then on the desktop system?

7. Parallel tools on a Ubuntu 9.10 server systems you can't install them.

8. You open a priority 1 mission critical ticket which they define as:

• Severity 1 (Urgent): A customer is unable to install and run the software because of problems with a license key, the computer (server) fails to start, or the software crashes and corrupts data. No workaround or immediate solution is available.

The next day (more than 13 h later) you still have not a single reaction while you sent like 10 mails and logs and reports and screenshoots. After 16 hours, the first reaction was, well it will probably be fixed in 2 weeks in the next release (maybe). If you operate a datacenter with thousands of servers as we plan it, you don't want to get into the situation that your servers reboot, get stuck and you have to wait 2 weeks until someone makes a working product.

Conclusion: There's no working product for the DataCenter from Parallels. If you thought about it, forget it. At least for the next 2 years. This looks like some russian cpu cracks who had fun writing a (probably pretty good) virtual machine but have forgotten what software quality means. If your business depends on it, you are better off with more expensive solutions like VMWare. The only problem there is that they have no solution for virtualizing MacOS X Server. Someone might shout, "what about Fusion?". Yes thats true but that's a desktop application. Its built for users sitting in front of a computer. You can not virtualize 100's of servers on a big fat CPU with many gigabytes of ram and disk that way (ever thought how 100 windows look like on your screen?). I'm sure VMWare could make a product that way which would even work but they don't see the commercial market for it yet.

So whats the only other option? Piling up MacMini's. One for every customer. Thousands of wires. Ugly cabling. Reliability questionable. Or piling up XServes. One for every customer. Reliable but expensive.

I'm dreaming of a working virtualisation solutions working on big iron's like seamicro running MacOS X Server. But there are obstacles:

- One is called Steve which doesn't want MacOS X Server to run on anybody else's hardware than his

- The other is Apple who doesn't produce big enough fat servers

- Tnd the third is Paralells which is unable to produce serious products

or VMWare who see's no market.

No wonder MacOS X Server is not common at all in the hosting compared to Linux. Even though its a very stable server system with a nice guy which adds a lot of ease of use for many people who are not day in, day out working on the command line.

Folks. Give us an option! There's business to grab here...

iPhone SDK nonsense

Dear Reader,

I've followed the iPhone SDK launch and I can only shake my head. So I decided to let free my thoughts about it.

Apple originally didn't wanted to have 3rd party Apps on the iPhone. Web 2.0 is all what you get. Bullshit said the developer community. The iPhone was hacked and thousands of more or less useful app's where created over night.

Well this is interesting as an experiment but for commercial developer use this is not solid ground because you can not safely assume everyone has a jailbreaked iPhone. At least it had pushed Apple to release an official SDK.

We have awaited the offical SDK with tension. Our goal was to solve the most burning problems of the iPhone in the european market which Apple has simply forgot to solve because they have no clue how advanced Europe is in using mobile telephony. We have identified the following hot spots:


- MMS
- Bluetooth modem
- Bluetooth filetransfer (OBEX)
- Bluetooth sync
- Accessing or importing adressbook from SIM card
- SCTP Support for applications.

While investigating those solutions we have faced the wall at all times.

1. You can build your application but you cant run it on the iPhone. I have nowhere seen how you get this "special" key and neither how you get and install it. A simple error -34 is all you are going to see.

2. A lot of things use reuseable code. Thats what dynamic libraries are there for. The SDK can not build any shared libraries. Ok you can copy all code into the application but thats not the idea. Our WAP stack project is killed because of this.

3. Background daemons are not permitted. So no daemon who listens to bluetooth filetransfers, synch's or functions as a modem. Bluetooth projects are killed because of this.

4. Kernel extensions are not permitted. This kills SCTP. A fantastic network protocol which would be perfect to be used on the iPhone. SCTP supports seamless handover between multiple IP connections. So you could start downloading stuff on WLAN and continue to do so over EDGE/GPRS without any interruption on application level. Linux has it, Solaris has it, HP Unix has it, Windows has it. MacOS X hasn't. There's a solution for desktop MacOS X but its kernel dependent as it has to use an unsupported API to install a protocol above IP (next to UDP and TCP). Apple has told me they will support it. Maybe in 10.4.8 but maybe only in 10.5. Well we are at 10.5.2 now and there is still no corresponding API which was promised at WWDC 2004, 2005 and WWDC 2006.

5. There is no API to work with SMS This means you can not receive SMS to be dealt with inside an application. So no WAP push or MMS notification can be handled! MMS project killed. Chess via SMS killed.

6. The sandbox does disallow us from using any hardware directly. So importing from the SIM card? Don't even think about it. Project killed.


Now what does this mean? My project for the 6 "most wanted features" have been killed. So I can start developing another Tetris and stay in line with 192764652 other developers doing the same. And of course i have to sell it at the same place as the others. How can a buyer decide which tetris to take? Well it means Apple will be the judge at the end of who gets business and who doesn't.

I consider the requirement to sell via iTunes store as pure offense. There are many european countries who dont evne have a iTunes store yet. There are most european iTunes store who don't have Videos or TV offers. It will take dozens of years until there will be a iTunes store in every european country being able to sell applications.

Also there's lots of other places people look for apps. Handango, Kagi or similar platforms. So why should I give 30% of my revenue to Apple so I'm locked into their platform where no one looks and my market is limited?

And while we talk about the market: If i look how many iPhones have been sold in Germany and compare the population of Germany to Switzerland, it would mean in the first 1-2 months there would be about 7000 iPhones sold if its launched in Switzerland one day. If I compare this figure to the information of Swisscom moble saying that they have 15'000 subscribers using an iPhone in their network, I must say that the market of the "official" iPhones will be the minority for a long time.

My personal opinion is the marketing manager for the iPhone should get kicked. He has no idea on how Europe works in regards to mobile phones. Adopting US views don't work (otherwise there should be 10x more iPhones sold in Germany and UK). Europeans are by far advanced and used to new mobile technologies. But that guy just hasn't got taken a look outside the borders.

Apple should come back to its roots and think of what made Apple big. It was the developers who made useful applications for their hardware. Today those developers are being treated like kids. The message Apple gives them is "Go play in the sandbox" (and pay Apple a fortune for the permission to do so).

I can only say, don't allow this to happen to you and start protesting.

Especially if you are not living in USA, you get kicked your ass many times. European developers pay 50% more for being a member of the ADC! Even ADC Premium members who pay a very high price don't get accepted by the iPhone SDK Developer program. Europeans get iPhones half a year later (UK, Germany, France), or not at all (all remaning countries). They don't get any useable SDK they could start developing. iTunes stores don't exist in some european countries.

There is only one way Apple will change its mind to abandon the closed community approach: LET THEM KNOW what kind of nonsense they produce. You can do that by:

1. Create bugreports / feature requests (http://bugreporter.apple.com) making it clear what you want from Apple. Make sure how you feel being put into a sandbox. Make sure you keep copies of the bugreports. Apple is known to delete them silently or make it invisible. ( If this happens, make sure they know you are upset about that. Also a very doubtful practice by Apple.


2. Swamp Apple Developer Technical Support with requests of missing SDK features. Ask them why europeans have no access to the required keys of the SDK. You won't get any answer but at least it keeps them busy thinking on what they could do wrong.

3. Print T-Shirts for WWDC 2008 with a good slogan like:

"Free Europe from Apple's slavery".
"Thank you Steve for for treating europe like shit"
"I'm Apple second class developer (non US) but I still write better code than you!"
"Free the iPhone"
"No censorship for iPhone Appications".

(who has better ideas?)


4. And if that all doesn't work there's two other alternatives.

- ignore the SDK and write no applications for the iPhone
- ignore the SKD and write App's for jailbreaked iPhones (Thats soon the bigger market anyway)


Bottom line:

Let Apple know what kind of nonse they have produced!