Simon Fell > Its just code
Simon Fell
Mark has a cool RSS discovery for backlinks tool.
Beyond Backlinks: Inspired by Jon Udell's recent articles, I've written a forward looking update to Manufactured Serendipity, focusing on new ways to manage the experience. [Sam Ruby] You've probably already seen it, but more good stuff from Sam.
Thursday, May 30, 2002
Joe seems interested in my idea so I played with Visual Studio’s Add Web Reference wizard to see how it handles link elements. It looks like it just looks at the first "alternate" link with type='text/xml'. This could be a problem for sites that want to link to both types of documents. [matt.griffith] This shouldn't be a problem going forward, as DISCO is dead, and WSIL [which replaces it], uses a meta tag instead.
Matt Griffith: Do any RSS aggregators use the HTML link element?
[via Joe Gregorio: Ideas, Ideas, Ideas] This sounds like a fantastic idea. Specifying the location of your RSS feed in a standard machine-readable format would solve a real problem. [dive into mark] Seems like a great idea to me, I've added it here.
Doug Kaye on ebXML ..... "IBM and Microsoft still aren't committed". Given that Microsoft aren't supporting SwA, seems like they don't care much for ebXML [which uses SwA]
Wednesday, May 29, 2002
Lots of hits for QuakerSoap in the UserAgents report, but can't find anything out about it, Google turns up nothing.
SOAPBuilders F2F next week, I'm plugging away at the new HTTP/1.1 transport for PocketSOAP, hopefully I'll have it wrapped up [or at least close] by then.
Tuesday, May 28, 2002
Leave it to Apple to make kernel panics look good. [Hack the Planet]
Monday, May 27, 2002
Spoofing URLs With Unicode [Slashdot: News for nerds, stuff that matters]
James posts the sample I was looking for, thanks. So there's no way to say it returns a particular SOAP message then ?
Anyone have a WSDL sample that shows how to describe a HTTP GET that returns a SOAP message ?
Congratulations to Sam Ruby who was just elected to the Apache Board of Directors.
Sunday, May 26, 2002
Anyone know how to get STLPort and CppUnit to play nicely together ?
No HTTP compression for the .NET Compact Framework, this is seems like a strange move, its seems to me that PocketPC based apps have the most to gain from supporting this.
HTTP digest authentication interop issues
Looking at some User Agent stats from soap.4s4c.com I see that IE is advertising what updates/patches are installed, e.g. Mozilla/4.0+(compatible;+MSIE+6.0;+Windows+NT+5.0;+Q312461;+.NET+CLR+1.0.3705), I'm guessing that the presense of the Q312461 part means that the patch detailed in this KB article is installed. Seems like a great way to enable some targetted attacks, any requests to a server from IE without Q312461 in the user agent string is vunerable. [note to self, use Opera 6 by default]
Saturday, May 25, 2002
COMUnit is a Visual Basic port of the JUnit framework that is designed to facilitate developing unit tests for Microsoft COM components and Visual Basic applications. Just started porting my somewhat ad-hoc collection of tests cases for PocketSOAP to COMUnit, I like it so far.
Cool, looks like my etag suggestion fell on reciptive ears, Joe has added this to Aggie, and DJ Adams has written an etag aware wget.pl
Friday, May 24, 2002
soap.4s4c.com now has a WSIL file
Bill Kearney has some comments on my earlier suggestion about aggregators using ETags. In response to Bills points, both Apache and IIS will automatically generate ETags for static files, IIS will also serve a Last-Modified header in this case as well [haven't checked this on Apache]. You only make a single pass, supplying the last etag in the next request, this requires you to store the etag value between agregator passes [as does an approach based on Last-Modified, If-Modified-Since]. In the specific case of RSS you'll propbably not see any issues using Last-Modified, If-Modified-Since rather than Etag and If-None-Match, however for HTTP/1.1 Etag is the preffered approach [particularly if you want to support resumable downloads]. In either case you need to store some additional data between requests, so I don't see that one approach is any harder than the other. How easy this is, largely depends on what HTTP client code you're using, as Joe points out this is pretty easy in .NET, I've also done this with libcURL, where its equally easy, YMMV.
There's a new build of the WSDL Wizard for PocketSOAP beta available.
Thursday, May 23, 2002
Just ducked into the local computer store to check out some Macs, those tiBooks are sweeeeeet.
I actually have two incomplete essays on the topic that I'm shelving as it seems that this is a topic is beyond rational discourse at the present time. [Sam Ruby] Sad, but true.
Ben Hammersley is collecting ideas for RSS evolution
Wednesday, May 22, 2002
SQL server worm throttles bandwidth. Spida spins a malicious Web [The Register]
However, Simon's answer poses another question: Why is SOAP the format/protocol to use with these kinds of protocols? I just gotta wonder why SOAP would be the wire format of choice over [reliable] mechanisms such as MSMQ/JMS, versus, say, other XML based formats, given SOAP's complextity in some areas... [Winterdom's Weblog] Well, let me ask this, why is SOAP the wire format of choice over HTTP ?
Tuesday, May 21, 2002
Tomas says Being a courious mind and all that, I ask: What other transports? Well, things I've seen code for include SMTP, Jabber, a simple TCP binding [implemented by SOAP::Lite and PocketSOAP], BEEP, MSMQ, JMS, and other message queuing products. One I haven't seen anyone support yet, but which I expect to gain some traction once more toolkits have DIME support, is the DIME/TCP binding detailed in the WS-Routing spec.
Jon asks Does SOAP routing necessarily preclude an HTTP-derived transport? It doesn't preclude it, but SOAP routing gets more interesting when you take HTTP out of the picture and look at other transports.
Webcasters win reprieve from fees. Federal regulators reject proposed royalty rates for online music broadcasts, which small Webcasters criticized as too high and the big record labels called too low. [CNET News.com] <vbg>
Joe's Aggie tool is OPML based and displays syndicated feeds in the browser using CSS and JavaScript to enable you to see the first line of every syndicated entry in full, and the remainder of the entry in a condensend font. To expand any entry, simply click on it. Clicking again toggles it back.
Full source, executables, and README are available as a 23K download.[Sam Ruby] Its neat, but to be fair, that 23k download requires the .NET runtime, a 21Mb download if you don't have it.
Monday, May 20, 2002
In researching some HTTP caching issues, I came across this interesting paper[pdf] : Clarifying the Fundamentals of HTTP
Sunday, May 19, 2002
Something funky going on with the aggregator in Radio, the Etag samples i posted choked the aggregator display, as far as i can tell, Radio doesn't encode the HTML when it generates the RSS, but the aggregator [or the news page] is doing a decode of what it got in the description tag. Perhaps Radio should be encoding the HTML tags in the RSS it generates ?
... Maybe Simon's willing to post some sample header values to shed some light on the idea... 
[Jake's Radio 'Blog] Sure, the server response may contain an ETag header, you save this away, ascociated with the URL, then on subsequent requests for that url, you include a If-None-Match header with the etag value. If the contents haven't changed, then the server will reply with a 304 Not Modified. I moved the samples to here, as they seem to choke the aggregator display in Radio.
Here's a suggestion for people writing RSS aggregators, use the HTTP/1.1 Etag and If-None-Match headers so that you only fetch the feed if its changed.
Woo Hooooo !, it finally stopped raining :)
I'm sure I've seen this mentioned before, but the ASP.NET webforms model is entirely POST based, I had to manually code up the column headers/sorting on this page, so that its GET/Bookmark friendly.
Following a discussion on axis-dev, I've put up a user agents summary page for soap.4s4c.com, so you can see what toolkits have been active in the last 7 days.
Charles Cook, the XML-RPC.NET guy has a weblog.
Saturday, May 18, 2002
Joe C. Carson: Turncoats. There have been a number of influential pro-Windows types who have been switching to Mac OS X, largely due to desperation borne of Windows' increasingly unfriendly behavior.
[dive into mark] I seem to recall Peter was thinking about getting a power book, which I didn't grok at the time, but I'm becoming more and more tempted myself to get something that'll run OS X.
Peter posted some answers. It seems to me that the GAC is as secure as the rest of the windows directory, i.e. good in theory, bad in practice.
Is Strong Named Assembly Really Secure? - Interesting, I thought that the test.exe manifest included the public key token of the version of real.dll that it got compiled against, and that this was verified at runtime. Where's Peter when you need him ?
The search for high level wsdl editing/generating tools continues: [John Burkhardt] I'm an Onmiopera fan, and the new beta has some great features like generating WSDL as two files, one for the abstract definitions, and a second with the binding in.
Friday, May 17, 2002
Jump/cut played last night at Platform, It’s got to be pretty rare to hear drum’n’bass performed with real drums and bass guitar rather than with a drum machine.
2600's DMCA Challenge Blocked. A federal appeals court thwarts, yet again, the hacker quarterly's attempt to have dismissed the decision that it is unlawfully distributing a DVD-descrambling utility. Declan McCullagh reports from Washington. [Wired News]
MS IE patch misses the mark. Partial fix leaves 12 outstanding bugs [The Register]
Thursday, May 16, 2002
Tonight - Platform - Showcasing new talent in design, film, fashion and sound.
Wednesday, May 15, 2002
Word Blogger is a port of my previous Word blogging macros to XML-RPC, this means they now work with other blogging tools besides Radio, and also you don't need to fiddle with Radio to enable the SOAP version of the BloggerAPI, it can all be enabled from the prefs page. It also includes a couple of tweaks suggested by Omar Shahine.
The next step is to call blogger.getRecentPosts so that its easier to edit existing posts from Word. The hurdle here is converting the HTML back into a sensible word document [surely Word must be able to do this for me ?]
Xbox Price Drops to $200 [Slashdot: News for nerds, stuff that matters]
Tuesday, May 14, 2002
SOAPBuilders is discussing dictionary serializations [again].
Been playing with MapPoint.NET. Just for grins, here's a little picture of roughly where my web site thinks you are - is it right? :-) [Peter Drayton's Radio Weblog] Neat!, for some reason though, it thinks I'm in Santa Cruz.
Monday, May 13, 2002
PocketXMLRPC v1.0 for PocketPC and Win32 released. Includes the PocketBlogger demo app for PocketPC.
Sunday, May 12, 2002
If you're a NSIS & COM guy, you might want to checkout this patch
More tcp.im thoughts from Eric. Jeremy posted notes on programming with Jabber. [Scripting News]
Matrox's New Three-Head Video Card [Slashdot: News for nerds, stuff that matters]
Last one for tonight
Saturday, May 11, 2002
another test.....
Ahh, found it
The Blogger API in Radio docs mention a "blogger.getRecentPosts" call, but this isn't listed in the Blogger API docs. This is a Radio only thing, or did it just not get added to the Blogger API docs ?
apparently so!
ls this thing on?
[powered by PocketBlogger]
Scripting and IM [Scripting News] Dave, any plans for SOAP over IM ?
More from Julian Bond on Web Services over IM. BTW, WS-Routing includes a standard way to manage Message-Ids.
Friday, May 10, 2002
I read this story tonight on Simon's site. Originally posted in March 2001. The good news is his prediction at the end hasn't come true. The bad news is in over a year there hasn't been a significant change to WSDL. It is still as complex, still as ambigious, and still tied to XML Schema. Although the last part I don't mind. [Justin Rudd's Radio Weblog] Well yes & no, AFAIK no one has even a remotely complete WSDL implementation, the current subset's that are being implemented seriously break some of the design points of WSDL [Messages aren't abstract because everyone ties the use of type or element to the binding style, so the whole concept of an abstract PortType is unworkable with today's tools]. Most depressing is watching the traffic on the WSDL WG mailing list, I don't see anything that indicates WSDL 1.2 is going to be any simpler than WSDL 1.1.
Debate foams over SOAP 1.2 [IDG InfoWorld] More SOAP vs REST, just in case you haven't had enough.
Movie Tickets. Hey, I just bought Star Wars tickets off of Fandango.com. Much better than what I did last time (wait in line overnight). [Robert Scoble: Scobleizer Weblog] Good luck!, the last set of tickets I brough through Fandango turned into a nightmare, I'll never use them again.
Must-read News.Com piece on MS's Web services strategy. 
[Scripting News]
BEA pushes new Java spec [IDG InfoWorld]
Thursday, May 09, 2002
... It is practically impossible for every implementation to implement every feature of the SOAP and WSDL specifications. It certainly can't happen "all at once," but only though a continual process. [deem] This is don't have a big problem with [other than its only complicated because you insist on bundling SOAP & WSDL together, SOAP on its own is certainly pratical to fully implement], but Keith has said that there is no commitment to fully implement section 5 in ASP.NET. This I have an issue with.
The new Omniopera beta has a cool re-authoring wizard, which splits a WSDL file into separate abstract & binding files.
Wednesday, May 08, 2002
Keith & Mike both seem happy to subset SOAP and WSDL. Wave bye-bye to the chance that SOAP interop means anything more than interops with MSFT. Where are all the MSFT guys who were up in arms 12-18 months ago about subsets ?
Care to enlighten us on what that subset is? Or are the ones that Keith mention the only features missing? [Justin Rudd's Radio Weblog] In addition to the items Keith mentioned, it doesn't handle the root attribute, partial arrays, or certain kinds of Generic Compound Types. Outside of section 5, header faulting is not handled correctly [it will generate a mustUnderstand fault, even if the header is not targeted at the ultimate destination]
Tuesday, May 07, 2002
Some pushback from Keith ... I wish we didn't have any section 5 encoding features. I wish you didn't have any section 5 encoding feature as well, rather than the subset you decided to implement. If you're going to provide section 5 support, then do it, if you don't like section 5, then don't support it, but providing a subset just confuses the issue.
More DIME interop, successfully tested against Bob's WhiteMesa DIME server. Interop results updated.
Mike talks about the MSTK, DIME and SwA. I find it ironic that the MSTK is billed as a fill in product and that .NET web services is MSFT's stratigic direction, given that in my experience the MSTK is significantly more compliant with the specs than ASP.NET web services.
Monday, May 06, 2002
Editors' Newswire for 6 May, 2002. Newswire stories, including: SOAP HTTP GET Binding. [xmlhack]
Sam goes onto say "What's more unfortunate is that the small gaffe obscured Keith's real question, which was how one determines the correct message definition associated with a given message. The answer is that you can't. A message may not have a definition. It may also conform to multiple definitions." I agree with this, but I also think it would be useful to be able to unambiguously refer to a WSDL described operation from outside of WSDL.
Sam saves me the trouble of responsing to Dare's comments, I guess the Infoset meme has taken hold more than i thought, I hardly ever think about namespace prefixs in XML, after all they're just a side effect of a serialization optimization. 99.999% of the time I say QName, I'm thinking of a namespace + localname pair.
Sunday, May 05, 2002
Read a bunch of good books recently, Ken MacLeod's The Sky Road, Cosmonaut Keep and The Cassini Division, Eric Nylund's Signal to Noise, Vernor Vinge A Deepness in the Sky, Greg Bear The Forge of God and John Barnes The Merchants of Souls. I'll probably switch tracks, and start reading Ingo's .NET Remoting book this week.
Well, that's it. I've given up CityDesk is favor of Radio, for one simple reason: remote posting. [The .NET Guy] Cool, Brad is migrating to Radio. Does this mean we'll loose the funky green & purple site ?
Saturday, May 04, 2002
Dave Winer, you are a very bad man!. [gRadio] ROTFL !
Anyone know if there's an XML interface to the song history in ShoutCast ?
... I'm going to spend some time looking at ATL 7, Boost, and Loki (from the book Modern C++ Design).[Justin Rudd's Radio Weblog] Cool, I'd be interested to know if Loki is usable from VS.NET, i tried it with VC6, and it fails miserable [because vc6 doesn't support all of the template bits]
Friday, May 03, 2002
Julian emailed me about the Word 2 Story macro, I've just posted a revised version that works with the newer versions of PocketSOAP.
Thursday, May 02, 2002
Keith Replies ... However, I don't think I'm understood. I don't mean, "how do I tell that a piece of XML is a SOAP message", but given any SOAP message, how I tell the "type" of that SOAP message? Sounds a lot like Tim & Martin's point that there's no common way to reffer to a binding/operation outside of WSDL [I guess some XPath type approach that works on the logical WSDL model is the closest you can get today]. I'm sure this'll be fixed in WSDL 1.2, right Keith ? But, if you're taking a messaging approach, then all you really care about is the schema right ?, for which you should be able to extract based on the QNames used in the message.
My proposal doesn't modify HTTP at all.... [Snell's Blog] There isn't a single HTTP stack out there that will know what to do with INVOKE. This negates 99% of the reasons to use HTTP.
....Given any XML document, you can look at it's default namespace and tell it's type. [KeithBa's Blog] I'm going to assume that this is a typo, if not, Keith should know better. The default namespace MEANS NOTHING! Its perfectly valid for a WSDL doc to not have the WSDL namespace as its default namespace (as long as the elements are qualified properly). The root element QName (the element local name + namespace URI) tells you that a WSDL doc is a WSDL doc, why is SOAP any different ?, the root element QName ( {http://schemas.xmlsoap.org/soap/envelope/}Envelope for SOAP 1.1 ) tells you that its a SOAP message.
I've organized some thoughts on the use of HTTP POST and GET with SOAP. After going back to the HTTP specification, I came up with a somewhat surprising (to me) conclusion: for SOAP RPC, what we need is a new HTTP Operation called INVOKE. Read this for details.
Bottom line: SOAP RPC SHOULD NOT USE HTTP POST.[Snell's Blog] If its going to take modifying HTTP to use it with SOAP, then it seems completely pointless to use HTTP at all, its hardly an ideal transport for SOAP to begin with.
Wednesday, May 01, 2002
Justin Writes ....I can also only assume that ASP.NET is discounted because out of the box it will match elements to parameters of the method. This is both fortunate and unfortunate. [News from the forest] I discounted it because by default it generate a <myMethod> element, although its not following the RPC part of the SOAP spec, it is still clearly RPC.
Cool ! Christian Forsberg has written a new PocketSOAP article Calling Secure Web Services for the Microsoft Mobile Devices site.
Last update: 2/9/2003; 5:31:46 PM.
