
Tuesday, July 24, 2007
Amazon Web Services Talk

Monday, July 23, 2007
Jeff Barr talk tonight
See everyone at 5pm tonight. Details here. We're at 61 registered and quite a few contacted me by email, so please do come along :)
Saturday, July 21, 2007
iPhone v Harry Potter
Wouldn't it have been neat for them to have lauched at the same time with Harry Potter available as an iBook on the iPhone.... what kind of queues would we have had then!?
But when it comes down to it, what we have is what you see below. A great couple of pictures for any lone entrepreneur starting with a new idea.
I wonder whether the iPhone bunch should have dressed up as mobiles....
iPhone
Harry Potter
Friday, July 20, 2007
Francisca and her first tooth
Meanwhile we got some old XBox games - first time we have bought any for years as we don't want Xavier obsessed with games. Right now he's playing Star Wars Lego edition.
Cool - we just managed to configure the Logica JoyStick i bought last week (which nothing worked with)... Xavi is loving it!
Talk on Monday
If you are in Scotland and have ANY interest in Web Services - make sure you DO NOT miss this presentation.
OpenID.ORG - *your thoughts* - PLEASE!!!
So I posted this response to Scott. I really don't feel any negative vibes toward them all despite the crap that has been said about me - probably because i have got a lot of support too and I only worry about my kids - other stuff doesn't concern me.
I want to hear from YOU though. Email me at "weblivz AT hotmail DOT com" with your thoughts.
I'm now wondering whether all the effort i have put in for things like mobile, microformats, foaf and so on has been worth anything. I have really enjoyed doing it, but this keeps coming up. Maybe it's time to sell up and get out - hard with a domain that was valubale before the OpenID concept ever came to fruition and which sits number 5 in Google - something I never intended doing... but i have too many other things in my life to worry about that to cope with people taking all the positive stuff i am trying to do and making it negative. With Google and archive.org in 20 years my kids will see all this stuff and wander what the heck i did that was so upsetting to a few people. The Microformats people said "cool". Some FOAF people said "neat". I have been working on some other ideas people may like... and it's taken me a lot of time. But this comes back again and again.
Anyway, by reply to Scott...
Let's look at this another way. It was actually useful for me to point at your site - you would never had asked me to do so otherwise. I got zero benefit from this. Remember, the original reason you asked me to do so was because people were ALREADY pointing at the .org rather than the .net - so the fact i redirected was to help you guys out. Had I not redirected I'd still be around the same place in Google - it wasn't *because* i redirected that this happened.
I am the villian to you guys for no reason other than i bought a domain name. Then you invent an idea that, had you thought it was to be as big as it is, you would have bought all domain extensions for it (i have done that with some of my ideas) and then get annoyed at ME for not handing it over.
I live in the real world where i have two young kids and a living to make somehow. So i put on some Google ads - I am now the proud owner of $8.95 this month. You guys offered me $2k - at the same time you offered a $50k bounty for applications on it. You also can also charge $100,000 for platinum membership of OpenID. You guys are also travelling around the world promoting this stuff - my last trip was 4 years ago. You all seem pretty well off by my standards. So imagine where i lose the domain name, and openid disappears and then two years later i see someone else using the OpenID.org domain for some identity system that is not the openid we know and love - perhaps it just never took off. You can't pretend to me that in that case someone hasn't made a killing on the domain name. May not happen, but that's technology.
You say "That’s stooping pretty low Steven." - man, I did that BECAUSE i didn't want people thinking my site WAS the real site. Honestly. It is to read that the openid.net is the REAL site and my is in no way affiliated.
And to say i am gleaning anything. I added mobile support, messaging, microformats and foaf (and some other stuff not yet released) to try and pull in some of the other thinking - it was to help the idea of OpenID... but you make it sound like these were already in place - as though i took some template of all the work people had done and dropped it in. I have spent countless hours working on this stuff (that's why i have trouble sleeping) - a big deal when you work full time and have two kids.
Trust me when i say i am raking nothing in because of all this. Really. I'd getting sick of all this though - maybe OpenID and all these well off sponsors, members and individuals should think together and make an offer i can't refuse. Not something I would have considered before, but i'm seriously getting frustrated by all this. But I can't really win in this. Ask around for what people think openid.org is worth - as a domain name in general and then with the OpenID concept around it - in both cases it's worth a lot as it's a cool domain. People who didn't even know there *was* and OpenID have said it's a very snappy domain. So whatever i could sell it for I won't get its real worth - i know that - espcecially if the foundation took it. But then if i sell it for a decent amount i'll be the villain of the community - i may even make Time Magazine and new releases of SpiderMan will do away with the Green Goblin and have me in its place (which may be pretty cool for my kids). It's not as though we were talking in millions of dollars - it's all the companies that are benefitting from this technology that talk in those numbers.
I don't know - anyone - please mail me at weblivz AT hotmail DOT com with your thoughts and ideas on this.
Thursday, July 19, 2007
Why Services will win over Installs and Source
My savior through all of this has been VMWare - i don't know what i would have done if i did not have snapshots!
However, it has made me start to think a little deeper about services (well, combined with some stuff i've heard privately and also online from people such as Jeff Barr and Don Box). It has driven into me how you doing software installations are on the way out. For personal users - and for business users. Especially the SME. Furthermore, Open Source projects are out too. Well, let me re-phrase before i get a million comments. Open Source projects that work will be distributed and managed through a small number of expert groups - rather than downloadble and distributed to everyone who wants the product.
Furthermore, IMHO, these Open Source projects will no longer be installable products - they will be API's. So will commercial software. Most software will. It's just toooo much hassle. A week wasted when i knew what i wanted - i really didn't want to know the details... just some services i could ask some questions of and get sensible answers back.
Take queuing (something i need soon, not yet). Amazon offers a simple queue service. Microsoft are looking at an Internet Service Bus. Others are doing similar work. These are accessible via API's and there is NO NEED for me to download and install this stuff. It's just too much hard work... well, when things go wrong it is - and my experience in IT is that if you have 2 complex things to install you're gonna have problems pretty quickly. I want people who create these services and are experts at that - they can scale them if i need and manage all the complexities around backup, failsafe etc...
I could write forever about these services such as authentication, storage, reporting, processing, e-commerce and so on. I want to drop these services as modules, combine them and add MY stuff on top (oh, maybe i should also write about modular UI AJAX/Flash/Silverlight style components that marry these services, such as authentication, registration, reporting etc). Writing and managing them all is just too much work. Importantly however, i don't want these as part of something like Facebook! I want them purely and simply as services. I know where i stand with an API which i know is a product and is supported as a core part of the business - not an addon. These services will need to stretch what we can do RESTfully and so SOAP wil come into its own.
Wednesday, July 18, 2007
The programmable web - mixed opinions
The issue is that they have been very pro-enterprise and less involved in the day to day hackers that are heavily involved in building many of the high profile web apps we see today.
However, i think a lot of the REST apps, driven by API's that use GET are medium term services..... on the way to what SOAP offers. I don't think SOAP itself is the problem, i think it's just that the people who would require the advantages that SOAP offers build more complex applications - they are the established companies, corporates and so on. People writing REST services won't want to manage session state, message level security and more - things that aren't easily avoided long term.
Many of the web api's out there using REST don't do a while lot in terms of message security, authentication, transactions or sessions. However, this is the initial wave - many of us ARE thinking of ideas requiring these services. There is also the evolution of web service interop to consider - a couple of years back i wrote some services that talk to Jave SOAP services and it wasn't easy. We did this internally on some basic ideas so could avoid some of the complexity of message level security and transactions and so on - but there was definitely thinking around scaling those services up - something only SOAP and not REST protocols would be fit for.
However most using REST are (like us) startups - larger scale enterprises are still figuring out what web services are and how they may be of value. They certainly aren't going to expose online transactional systems on the web via simple REST API's. Some may try, but most won't. So REST is a good start and no doubt will fit many scenario's but i firmly believe that as services get more complex SOAP will start to become a more popular format... i just cant imagine all my data being exchanged as hacked versions of RSS/ATOM :)
But I spent 4 years writing Perl code in the early 90's and everything worked great - until the time to build, tooling, scalability of the compiled langauges offered so much it was just too hard to resist. Microsoft just need to make the time to build simple REST apps as quick as the other toolsets out there - they have the complex, scalable stuff covered.
Oh, i'd like to see an Xml-RPC intergave to WCF services if possible!
Tuesday, July 17, 2007
Window Communication Foundation Unleashed - Xml Fetish
I was sent a copy as i made a comment some time ago on integration of Xml Schema constraints with Web Services. I read his argument in the book, but i cannot agree it is the right way to do things.
His argument is to convey the business rules (well, the basic rules initially) through the name of the property. So you may have a property called "QuanityBetween10And1000" rather than define a simpleType in your schema that states that "Quantity" must be a value between 10 and 1000.
His argument is at its core based on the fact that current proxy generation tools (such as Microsoft's xsd.exe) do not go any deeper into the schema than the name and type of the property. What happens if you have two rules, or three?!
IMHO this is a tooling issue. If i create a web service interface and set a simple constraint (complex types are more difficult) then i expect good tools to take this into consideration - i really can't imagine why not!? Especially with the relative overhead of a web services call, the more intelligence that can be put into a proxy generated class, the better... and Xml Schema is cross-platform and cross-language. So ANY language can get these rules. if the proxy generation tool doesn't support it then fine, they get the hit. But most really should.
In fact, in WCF i'd expect a proxy generated class to not only provide a typed and named property, but i'd expect any simpleType definitions to be included. How hard would a check on an integer range be? And it adds a lot of value. Something like BPEL will do this with even more sophistication... so why not put in the beginnings of this.
The book is excellent, but as the point above is probably the reason i got a mention in the book in the first place, i felt i had to write something. I Craig doesn't ask for it back :)
Friday, July 13, 2007
Facebook for Single Sign On
I was reminding of this by a recent twitter post (it hurts calling it a "tweet"!!):
I predict that Facebook will release an "Identity/Login widget" and will
overtake Open ID, why? Consumers rule, and Open ID is too geeky
I don't really agree that FB will overtake OpenID and then become some kind of central identity or authentication mechanism. Something we already have seen tried with MS Passport (where Jeremiah sites the trust issues with MS).
Now, i'm not specifically referring to OpenID [ i approached Scottish Enterprise way back in 1998 about an open approach something open, but similarto MS Passport ] so i'm talking about a general authentication mechanism for the web if you like...
My reasons are:
1. This should be a service. My view is that companies providing and managing this kind of data should be a service which doesn't directly conflict with half it's customer base. If you offer identity and then a bunch of web apps and api's, then it makes it harder to convice others that you won't simply lock them out by stopping them accessing the data you hold (even if it came via that site in the first place).
Ironically, this Twitter post from VC Fred Wilson was quotes Mark Zuckerberg, founder of Facebook:
'being a tech company means you aspire to be a platform, a layer in the stack
that others can build on'
Yes! But i can't see that happening when you ARE the stack. Facebook is cool, but we need to be careful what constitutes an identity service layer. Do we want to mix that (would TCP/IP work as well if it was the OSI model, rather than two layers within it)?
2. FB is a commercial company with fingers in many pies - other companies will not just hand them their most important data. It's going to be hard enough for OpenID providers to accomplish this - but at least they are somewhat independent. I already talked about FSBSoftware who were told to remove their sychronization application..... the user should be the one who controls their own data and where it sits.
3. Decentralization. There are too many scenarios in the real world where you, or a company you work for doesn't want to use an external site. OpenID means writing your code once and interacting with external bodies easier.
I also don't think OpenID is too geeky, but that's kinda like me saying that Scotland or Chile are brilliant places to go holiday.... but a username, password and email are all you need.
Wednesday, July 11, 2007
Jeff Barr - Glasgow Amazon Web Services Talk
Click here to get all the information about this event.
Also join the Google discussion group at
http://groups.google.com/group/techscotland
Monday, July 9, 2007
FOAF support in OpenID
A short while ago I saw a tweet from blognation about OpenID and FOAF support. I was going to mail and ask how this should work, but decided to go ahead and implement it (sacrificing my lunch I should add).
So, thanks in part to Johnny Cash (playing in background) I now have an implementation that now works with anyone who has an OpenID profile over at http://openid.org
An example is over at http://steven.openid.org/foaf for my openid http://steven.openid.org
That makes around 1800 people who can now use FOAF with their OpenID identity. Hopefully some people will play around with it. I'm glad to see this as i'd using FOAF a lot a few years back. I'd REALLY pleased to see these technologies starting to talk to each other - that is the most exciting part of it all.
Now I'm away for some soup as i'm starving!
Sunday, July 8, 2007
A look at the past
hotmail.com
December 21, 1996
gmail.com
December 23, 1996
socialtext.com
May 20, 2002
sun.com
October 20, 1996
ebay.com
June 14, 1997
yahoo.com
October 17, 1996
scripting.com
December 27, 1996
xmlrpc.com
mentioned as i get a mention on the front page.
What the heck was SOAP anyway!?
Feb 29, 2000
Unfortunately i can't find out much about "Soliton Computing" - a company i started with a then friend back in 1992/1993 (doesn't seem to be indexed anywhere anymore :( ). We were running our business off the Univesity computers and they we're best pleased when they found out. In all honesty we were just creating web sites for aroudn 50 small businesses in Glasgow - the whole business part of it wasn't something we thought about. Shame. We just enjoyed hacking html, Perl and so on.
The oldest house in Glasgow
Glasgow has a load of interesting things to see and a couple of hours turned into 8 hours (it's on the other side of town which we don't visit too often).
Friday, July 6, 2007
Google banned from discussing Gmail
While i see Google's issue (and this is the first i have read of this), i'm with the gmail.de guy. It is his domain name. You can't just go deciding you want to start a service and so anyone with the domain name is breaking the rules. Seems there are cases pending in other countries and i suspect they'll all go the same way.
Do ratings matter? A completely unscientific look at YouTube
The issue seems to be that so few people actually rate or make a comment and the percentage that vote seems to be fairly constant no matter how many people view a video. What is interesting is that for videos involving an individual doing something visibly difficult (such as in sports) more people seem to vote and comment (again more analysis would need to be done but i looked at quite a few beyond that below and it was fairly consistent).
Now, you can argue that the point of user generated content is that the Wisdom of Crowds effect means that small numbers of the crowd voting builds an emergent validity to the data. So in Wikipedia few contribute relative to those who read. However, it's much easier to click to rate or make a throwaway comment on a youtube post (i'd say 1/3rd of these comments are arguments!) - therefore i'm not so convinced that what works for Wikipedia (where you typically need to have some level of knowledge about the subject) will work for other user collaborative systems.
I don't know who is rating, but may of the comments are argument between members who disagree with what has been said. Not sure if you can rate multiple times.... but if 1 in 500 random people made a comment on my music tastes i'm not sure it would really mean anything to me. The problem is that you really need to have some knowledge of the background of that person before you can guage the value of the comment (if that 1 person was Brian Eno it may make a difference to me!). In a single click to rate, the meaning is almost pointless to a flaw - i'm not sure rating should really even be in there. Perhaps when something has been very highly rated by a high % of people who view it there may be something in it, but otherwise it really skews what you look and - and probably what others look at too. If only 1 in every 250 people with "bad taste" (relative to me - imagine the other 249 have "good taste") vote for a pointless video, it will be one of the top videos on YouTube.
Anyway, here are a few measurements i made on viewings in YourTube. Notice that sports videos that capture individuals doing cool things are as popular as Larry Page and Steve Jobs videos... and us tech folk are well versed in rating systems and form a fairly niche market.
Larry Page speaks at the AAAS
http://www.youtube.com/watch?v=8_3OCq_vTWM
24,036 Views
95 Ratings
43 Comments
0.004% (1 in 250)
Steve Jobs Stanford Commencement Speech 2005
http://www.youtube.com/watch?v=D1R-jKKp3NA
604,988 Views
2493 Ratings
410 Comments
0.004% (1 in 250)
Coldplay - Yellow
http://www.youtube.com/watch?v=qI8I6qcxWyU
1,491,143 Views
4238 ratings
1756 Comments
0.003% (1 in 333)
U2 - One
http://www.youtube.com/watch?v=JFWPeVfWB9o
712,553 Views
1429 ratings
422 Comments
0.002% (1 in 500)
Perfect Day
http://www.youtube.com/watch?v=GrGT-CSmBZg
686 Views
2 Comments
2 Ratings
0.003% (1 in 333)
Carlos Santana & Mana
http://www.youtube.com/watch?v=GrGT-CSmBZg
118,134 Views
188 Ratings
65 Comments
0.0016% (1 in 625)
Featured items...
Obama Girl vs...McCain Mama? (#6 most subscribed)
http://www.youtube.com/watch?v=86HEv_Wtyj8
25,104 Ratings
65 Ratings
69 Comments
0.0026% (1 in 386)
Speed Stacking - Nearly World Record (first in sports category)
http://www.youtube.com/watch?v=U951R_r-3fM
73,805 Views
549 Ratings
1323 Comments
0.007 (1 in 134)
NBA Finals Lebron Etch A Sketch (sports cat)
http://www.youtube.com/watch?v=nYM__s3R5q0
1,464,610 Views
7442 Ratings
7543 Comments
0.005% (1 in 197)
Wednesday, July 4, 2007
WCF interop with complex data types
function SimpleRequest()
{
var xmlhttp = new XMLHttpRequest( );
xmlhttp.open ('POST', 'http://localhost:2407/TestInterop/Service.svc/', true);
xmlhttp.setRequestHeader('SOAPAction', 'http://tempuri.org/IMyService/MyOperation1');
xmlhttp.setRequestHeader('Content-Type', 'text/xml');
var data = '';
data += '<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">';
data += ' <s:Body>';
data += ' <MyOperation1 xmlns="http://tempuri.org/">';
data += ' <myValue1>Steven</myValue1>';
data += ' </MyOperation1>';
data += ' </s:Body>';
data += '</s:Envelope>';
xmlhttp.onreadystatechange = function ( ) {
if(xmlhttp.readyState == 4) {
alert("status = " + xmlhttp.status);
alert("status text = " + xmlhttp.statusText);
alert(xmlhttp.responseText);
document.form1.result.value = xmlhttp.responseText;
}
};
However when using complex types, things are not necessarily more difficult, but there are some things to watch out for. Specifically, you need to ensure that the SOAP request you build correctly sets the namespace of the complex type you are sending to the WCF service. So if you have defined a DataContract as follows in the service:
[DataContract(Name = "DataContract", Namespace = "http://livz.org/wcf/name")]
public class DataContract
{
string firstName;
string lastName;
[DataMember]
public string FirstName
{
get { return firstName;}
set { firstName = value;}
}
[DataMember]
public string LastName
{
get { return lastName;}
set { lastName = value;}
}
}
Then you must ensure to set this namespace in the request. In fact, if you do not, a default namespace ("http://schemas.datacontract.org/2004/07/") is applied to the complex type and simply passing the serialized data will not work (well, it does not fail, but the values of your type are not set!). So, to correctly make the request for the above data type, you can use the following JavaScript:
function ComplexRequest()
{
var xmlhttp = new XMLHttpRequest( );
xmlhttp.open ('POST', 'http://localhost:2407/TestInterop/Service.svc/', true);
xmlhttp.setRequestHeader('SOAPAction', 'http://tempuri.org/IMyService/MyOperation2');
xmlhttp.setRequestHeader('Content-Type', 'text/xml');
var data = '';
data += '<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">';
data += ' <s:Body>';
data += ' <MyOperation2 xmlns="http://tempuri.org/">';
data += ' <dataContractValue xmlns:a="http://livz.org/wcf/name"><a:FirstName>Francisca</a:FirstName>
<a:LastName>livingstone</a:LastName></dataContractValue>';
data += ' </MyOperation2>';
data += ' </s:Body>';
data += '</s:Envelope>';
xmlhttp.onreadystatechange = function ( ) {
if(xmlhttp.readyState == 4) {
document.form1.result.value = xmlhttp.responseText;
}
};
xmlhttp.send(data);
}
Tuesday, July 3, 2007
Partnering on Vidyo
Now either people do this or it is done programatically. The latter could be done by assigning everyone and OpenID and parsing the RSS feeds from the various media broadcasting sites and updating the site. I may do that.
The other option, is to get the various sites to "ping" the site with updates. I am trying to speak to some of the video broadcasters to see whether this is possible via the new API. This would be neat as i could get instantaneous updates of broadcasts. After all, the idea is about immediately upcoming broadcasts rather than scheduled broadcasts, so the RSS option doesn't work so well (they'd quickly get pissed if i kept hitting their sites!!).
So, an email has been sent to some of the big players in the hope some, or all get back to me. Of course they may say why on earth would we want to mix syndication with our competitors, but that - for me - would sure be nicer than the equivalent 800 social network feeds that i get today. It would also be a nicer user experience for broadcasters (one place to update and share) and consumers (one place to subscribe).
We'll see how things go ......
Vidyo.TV Launch
Still, at this point it's better than the alternative (nothing) so it's quite neat to play around with. PC webcasting will still be ahead of the mobile equivalent for a while i think.
Monday, July 2, 2007
Stone of Destiny filming
However we did see Robert Carlyle (who bizarrely lives not far from the filming at all, but still got a car...probably not eaten for the day and going somewhere rather than popping home and raiding the fridge :)) and the director Charles Martin Smith. Robert Carlyle is actually one of my favourite actors - mainly for the reason that we i saw the launch of Trainspotting here in Glasgow, he was the only actor who was (for me) better than what i had read in the book. There were loads of others who i'll likely come to recognize, but my son was more interested in the guys carrying the lights! He thought they were the key guys to the whole thing. Fantastic :)