You can subscribe to this list here.
2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(7) |
Dec
(45) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2003 |
Jan
(16) |
Feb
(23) |
Mar
(62) |
Apr
(33) |
May
(35) |
Jun
(37) |
Jul
(45) |
Aug
(15) |
Sep
(22) |
Oct
(41) |
Nov
(23) |
Dec
(17) |
2004 |
Jan
(14) |
Feb
|
Mar
(55) |
Apr
(8) |
May
(1) |
Jun
(11) |
Jul
|
Aug
|
Sep
(20) |
Oct
(11) |
Nov
(10) |
Dec
(14) |
2005 |
Jan
(4) |
Feb
(2) |
Mar
(6) |
Apr
(26) |
May
(5) |
Jun
(11) |
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2006 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
(18) |
Dec
(40) |
2007 |
Jan
(7) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Asynch M. <asy...@ho...> - 2003-02-08 22:11:35
|
----- Original Message ----- From: "Adam Rifkin" <ad...@xe...> > First off, there's no KNN anymore. We should get modpubsub.com up and > running soon, to serve as the *new* KN Network. (Yes, with a weblog, > too. Joyce just loves it when I volunteer her for things. :) Would you like to use xmlrouter.net? I also have topiczero.com & some others... |
From: <ad...@xe...> - 2003-02-08 16:42:26
|
Derek wrote: > Thanks Kragen! Sorry for the tardiness in responding, it's gotten kinda > busy round here. Whew, I think we all know what that feels like nowadays. > I forwarded your comments on server-push v. client-pull > to the back-end architecting (gahh, when did 'architect' become a verb?) > folks at Ludicorp (vide., <http://gameneverending.com/>), it was exactly > the sort of information people will most need to know when deciding what > technology to base their massively multiplayer online game (for example) > on. And so why aren't such figures published front-and-center on the KNN > site?? (Not-so-subtle nudge-nudge for KnowNow employees on the list.) First off, there's no KNN anymore. We should get modpubsub.com up and running soon, to serve as the *new* KN Network. (Yes, with a weblog, too. Joyce just loves it when I volunteer her for things. :) Fyi, I run a local copy of mod_pubsub's kn_apps on http://pubsub.knownow.com/beta/kn_apps/ A lot of the apps are undocumented or broken, but that's just the way it is. They're all on my "to do" list to fix. Anyway, thanks for the PHP bits below. I'm in the middle of trying to build a new Python Microserver, but PHP connectivity is high on my list of things to work on, too... ;) Adam The rest of Derek's message: > Okay, I dug up the PHP bits, it's so simple it hurts -- just gotta watch > out not to rely on ob_implicit_flush(), apparently it's still unstable so > use an explicit flush() instead. Some sample code (wch. might not survive > Yahoo-mail, they're getting testy abt. inline script stuff), e.g.: > <?php > function msleep($msec){ > $start = gettimeofday(); $usec = $msec*1000; > do { $stop = gettimeofday(); > $timePassed = 1000000 * ($stop['sec'] - $start['sec'])+ > $stop['usec'] - $start['usec']; > } while ($timePassed < $usec); > } > function send($msg) { > $pre = "<script>parent.msg('"; // start JS call to parent > $post = "')</script> "; // wrap-up the JS call > echo str_repeat(" ", 300) . "\n"; // MSIE caching fix > echo "$pre $msg $post" ; // buffer the message > flush(); // send the packet > } > // set_time_limit(5) ; // optional 5 sec. timeout (default = 30 secs) > $z = 10; > while ($z>0) { > msleep(1000);> > > > // Sleep 1 sec. > send("<h1>".$z--." ... </h1>"); > } > msleep(1200); > send("<h1>... shooka-pata-tata, hey!</h1>") ; > ?> > While the client page'd look something like -- > <head> > <script> > function msg(content) { > status = " "; // suppress URL display in status bar > document.getElementById('display').innerHTML = content ; > } > </script> > </head> > <body> > <div id="display"></div> > <iframe style='display:none' src="pusher.php" name="pushlet"> > </body> > So I guess you'd just round-robin through the currently connected client > IP addresses any time there's something to send 'em, and like that ... |
From: Adam R. <Ad...@Kn...> - 2003-02-08 16:32:59
|
I'm forwarding the message below in case anyone's interested in looking = at Windows scripting. I have a tarball that accompanies Andrew's = original message if you're interested. Also, my management wants me to recruit more developers and/or encourage = more downloads. Before I start randomly signing up Sourceforge = accounts, if there's anyone you know of who might be interested, please = let me know and I'll sign them up... you know how shy I am about = spamming lists with "C0M3 S33 MY N3W W4R3Z, D00DZ..." :) Adam -----Original Message----- From: S. Mike Dierken [mailto:mdi...@ho...] Sent: Thu 2/6/2003 10:30 PM Subject: Re: Windows Scripting I think the WScript approach is beautiful. There are a lot of things you = can do with it. In fact, I write my 'sensors' with it & add them to a = schedule in Win2K. Doing a GET of some RSS, parsing the XML and POSTing the items = is really simple. > I have also attached a version of the same script written in VBScript (pubsub.vbs). I love scripting, but VBS?... yuck. > Ben: Think about what a JScript microserver means...? (the current JavaScript MS assumes > an IE context, how different would it be to assume a WSH context?) You can probably build wrappers around the script host things like CreateObject, alert(), etc. But the biggest problem will be the object model - the COM interface is = not exposing the same one as the JScript client. Methods names might be = similar, but that's not the whole story. For example, with COM you have event.Parameter("name") =3D value; but = with JScript you have event.name =3D value; - so much smoother... The ability = to do this comes when you implement IDispatchEx (or something like that...). = The obj.prop syntax gets translated into a getProperty("name")/setProperty("name") method automatically - but you = have to implement the IDispatchEx interface. You should be able to find = examples on the Net - if not I have working code here somewhere.... Also, = requiring a KNEvent object as a parameter to publish() is different than accepting = JS Object - you might want to accept IDispatch and branch on the interface type - that way you can accept JS objects as well as the = custom/proprietary KNEvent ;) > I don't like VBScript, but if I have to use it I will. Good man. > Scripts can be used to quickly build COM objects (known as > Windows Script Components). I've heard of this but not done it myself. ----Original Message by Andrew: All of us have probably long suspected that the Windows Scripting Host = and Windows Scripting Environments might be ripe ground for building = quick demos and/or tests--especially in combination with the WMI = (Windows Management Instrumentation) capabilities built into Win2k. =20 I have recently shed my phobias and learned what I needed to know about = Windows Scripting. God I hate reading Microsoft documentation. 18,000 = link-traversal and concept-stack-pops later, I am able to offer you the = following example: =20 //pubsub.js =20 var shell=3DWScript.CreateObject("WScript.Shell"); var knms=3DWScript.CreateObject("KNMicroserver.KNMServer","knms_"); var ev=3DWScript.CreateObject("KNMicroserver.KNEvent"); =20 knms.ServerURI =3D "http://localhost/kn"; knms.Subscribe("/what/foobar"); =20 ev.Parameter("kn_payload")=3D"hi I'm here"; knms.Publish("/what/foobar",ev); shell.PopUp("Click to exit..."); =20 function knms_OnEvent(ev){ shell.PopUp(ev.Parameter("kn_payload")); } =20 =20 This example requires the windows microserver, but nothing else to = create it or run it (Windows Scripting Host is an integral part of Win2k = and also Win98 I believe). You can just double click it to run it. =20 I have also attached a version of the same script written in VBScript = (pubsub.vbs). Ben: Think about what a JScript microserver means...? = (the current JavaScript MS assumes an IE context, how different would it = be to assume a WSH context?) =20 You'll also find publish_processes.js, which is a JScript that publishes = all known information about processes running on the current machine to = the /what/processes topic on the localhost router. Note that WMI = scripting (of which publish_processes.js is an example), seems to be a = LOT easier in VBScript rather than JScript. I don't like VBScript, but = if I have to use it I will. See the VBScript_vs_JScript pdf. For more = information on WMI, check out http://msdn.microsoft.com and search for = "WMI" and "WMI Scripting" =20 Our PocketNOC example could be written as a script, thereby simplifying = it considerably. =20 There is an event-based interface for WMI as well, which I suspect will = allow for "live" monitoring of activity rather than polling. WMI also = allows for system control (e.g., termination of processes, etc.). One = could build a senctl-like interface (see gif file) for windows, rather = than UNIX, using WMI. =20 Scripts can be used to quickly build COM objects (known as Windows = Script Components). A loadable filter module that knew how to attach = and call out to a named COM class would then allow for arbitrary windows = scripts to be invoked for filtering/transformation in the router--slow, = but scripted and familiar to windows hackers. One could also make the = router module a script host directly, thereby avoiding slow COM = invocations--I have an article outlining the steps. |
From: <ad...@xe...> - 2003-02-08 16:28:06
|
Mike wrote: > This demo that Jon talks about it very similar to KnowNews - streaming > news items with discussion on the side. The discussions aren't saved > anywhere, but, like, that, what, is so easy, dude, come on. Thanks for the reminder about KnowNews, I linked to it from the kn_apps/index.html page and checked the new index.html in. I'll have to make a new tarball soon, I keep committing small but useful changes. Make sure if you have a CVS checkout to regularly do CVS updates so you stay... er... up to date. For example, do_method=whoami was added to pubsub.py ... of course, when running pubsub.py I see that clens does not, at present, work. Will need to go through it carefully to determine why. > The only thing this needs is a service that emits RSS news items - > this is a situation where doing a GET can create or start processing, > yet can still be safe and idempotent. Good point. > You do a GET with the URI being the service combined with the RSS feed > you want & this kicks off the polling monster to start periodically > fetching RSS and returning it in a mod_pubsub stream (if it isn't > already doing it for another client). When there are no more clients, > then you stop. The new check-in of cxx_pubsub, among other things, contains mod_pubsub/cxx_pubsub/LibKN/Apps/RSSPub which could be used to create such a service. It uses the new LibKN (sorry Fred, there are multiple libkn's now!) available in cxx_pubsub, based on a C++ framework that has the following features: * Supports all Windows platforms: Win98, WinME, WinNT, Win2K, WinXP, and WinCE 3.0. * Multiple interfaces to the C++ code: COM/ActiveX and .NET. * Thread-safe supporting multithreaded applications. * Doxygen comments. * cppunit-based unit tests. * Sample applications. * Based on WinInet to support proxies and SSL easily. Check it out, you won't be disappointed. > You could alternatively POST to a special service with the RSS > location and the mod_pubsub destination, but you can't code that in a > URL that you pass to kn.subscribe(url) Another good point. Mike's on a roll. :) ---- aD...@Xe... Websites Complaining About Screen-Scraping Posted by michael on Friday February 07, @03:18PM from the there's-always-a-whiner dept. wilko11 writes "There have been two cases recently where websites have requested the removal of modules from CPAN. These modules could be used to access the websites (EuroTV and Streetmap) from a PERL program. The question being asked on the mailinglists (threads about EuroTV and about Streetmap) is 'can companies dictate what software you can use to access web content from their server?'" -- http://yro.slashdot.org/yro/03/02/07/1849243.shtml?tid=156 (as pointed out by the ever-knowing Greg Burd... :) |
From: Asynch M. <asy...@ho...> - 2003-02-08 07:12:04
|
This demo that Jon talks about it very similar to KnowNews - streaming news items with discussion on the side. The discussions aren't saved anywhere, but, like, that, what, is so easy, dude, come on. The only thing this needs is a service that emits RSS news items - this is a situation where doing a GET can create or start processing, yet can still be safe and idempotent. You do a GET with the URI being the service combined with the RSS feed you want & this kicks off the polling monster to start periodically fetching RSS and returning it in a mod_pubsub stream (if it isn't already doing it for another client). When there are no more clients, then you stop. You could alternatively POST to a special service with the RSS location and the mod_pubsub destination, but you can't code that in a URL that you pass to kn.subscribe(url) == http://weblog.infoworld.com/udell/2003/02/05.html#a596 Wednesday, February 05, 2003 Exploring Tim Knip's Radio/Groove interop tool When you install Tim Knip's Groove/Radio interop tool you specify three tools in a shared space. First, a discussion tool whose items are sent to your blog. Second, a files tool whose contents are upstreamed. Third, another discussion tool that collects the RSS items fetched by your news aggregator. That third step is a doozy. Last June, in a report on an experiment in using Groove and weblogs together, I wrote: Hugh Pyle added something that made the experience truly Groovey: an RSS reader. Suddenly the experience became qualitatively different. This news aggregator was a group resource. I immediately saw it as a way to work more effectively with (for example) my new colleagues at InfoWorld. I know of no other way to focus the attention of a group on a stream of news which is guaranteed to be identically and persistently available to everyone, and at the same time to be able to support collaboration around that stream -- i.e., discussions about which items to pursue. Tim's RSS-to-Groove feature accomplishes the same thing. But for me, and for most people, it does so in a much more accessible way. Hugh's implementation was a native Groove tool. Writing one of those is dramatically simpler thanks to Groove's latest toolkits and APIs, but it's still a fairly daunting exercise. With Groove Web Services, Tim was able to write his version in Radio's UserTalk. Very exciting! |
From: Adam R. <Ad...@Kn...> - 2003-02-07 00:43:19
|
I added the "Options +ExecCGI" to the INSTALL file step 5, because my Apache Server was hiccuping without it. Fyi, Adam 5. Configure your Apache server to allow scripts in mod_pubsub/cgi-bin to be parsed by mod_perl, and to alias the URL "/kn" to pubsub.cgi. You can do this by adding lines like the following to httpd.conf. (You may need to adjust the filenames for your Apache document root.) <Directory /var/www/mod_pubsub/cgi-bin> <Files *.cgi> Options +ExecCGI SetHandler perl-script PerlHandler Apache::Registry PerlSendHeader On </Files> </Directory> Alias /kn /var/www/mod_pubsub/cgi-bin/pubsub.cgi |
From: Adam R. <Ad...@Kn...> - 2003-02-06 07:08:26
|
I really like the way he thinks... http://www.oreillynet.com/lpt/wlg/2572 A Safari of Net Effects Part 2. by Timothy Appnel Jan. 9, 2003 In the continuing the review and discussion of Safari, Jason Kottke asks = the intriguing question why are Safari and Sherlock two different = applications? Jason argues that there is little distinction between web = browsing and using specialized interfaces for structured data. He = provides screen mockups of Safari to illustrate his point. An active = discussion in the comment boards follows. Back in September I wrote as the Internet continues to evolve into an = 'Internet operating system'--programmable interfaces, ubiquitous access, = and distributed computing resources--the document-centric browser is an = awkward solution to a growing number of emerging needs. The browser is = not dying by any means; it just needs a mate. Reading about Remote Application Development with Mozilla, the = mod_pubsub open source project KnowNow kicked off and discussions like = the one Jason is leading have me reconsidering my view. Does the browser = really need a mate in as much as it needs to expand its range? Could browsers like Mozilla and Safari/Konqueror be the basis of simple = lightweight structured interfaces for accessing network resources and = microcontent? What if these browser brought bookmarklets and remote XUL = to the forefront as equal partners to viewing webpages? More intriguing questions and experiments lie ahead. Timothy Appnel is an independent consultant and writer specializing in = emerging technologies and trends.=20 |
From: Adam R. <Ad...@Kn...> - 2003-02-06 06:56:26
|
With this release, we have added: 1. An index page for kn_apps. 2. C/C++ pubsub client (c_pubsub and cxx_pubsub) for Posix and Win32. Also, we started the transition to /kn ... Download from: = https://sourceforge.net/project/showfiles.php?group_id=3D66293&release_id= =3D138350 Enjoy, Adam |
From: Derek R. <nn...@ya...> - 2003-01-29 17:46:42
|
Thanks Kragen! Sorry for the tardiness in responding, it's gotten kinda busy round here. I forwarded your comments on server-push v. client-pull to the back-end architecting (gahh, when did 'architect' become a verb?) folks at Ludicorp (vide., <http://gameneverending.com/>), it was exactly the sort of information people will most need to know when deciding what technology to base their massively multiplayer online game (for example) on. And so why aren't such figures published front-and-center on the KNN site?? (Not-so-subtle nudge-nudge for KnowNow employees on the list.) Okay, I dug up the PHP bits, it's so simple it hurts -- just gotta watch out not to rely on ob_implicit_flush(), apparently it's still unstable so use an explicit flush() instead. Some sample code (wch. might not survive Yahoo-mail, they're getting testy abt. inline script stuff), e.g.: <?php function msleep($msec){ $start = gettimeofday(); $usec = $msec*1000; do { $stop = gettimeofday(); $timePassed = 1000000 * ($stop['sec'] - $start['sec'])+ $stop['usec'] - $start['usec']; } while ($timePassed < $usec); } function send($msg) { $pre = "<script>parent.msg('"; // start JS call to parent $post = "')</script> "; // wrap-up the JS call echo str_repeat(" ", 300) . "\n"; // MSIE caching fix echo "$pre $msg $post" ; // buffer the message flush(); // send the packet } // set_time_limit(5) ; // optional 5 sec. timeout (default = 30 secs) $z = 10; while ($z>0) { msleep(1000); // Sleep 1 sec. send("<h1>".$z--." ... </h1>"); } msleep(1200); send("<h1>... shooka-pata-tata, hey!</h1>") ; ?> While the client page'd look something like -- <head> <script> function msg(content) { status = " "; // suppress URL display in status bar document.getElementById('display').innerHTML = content ; } </script> </head> <body> <div id="display"></div> <iframe style='display:none' src="pusher.php" name="pushlet"> </body> So I guess you'd just round-robin through the currently connected client IP addresses any time there's something to send 'em, and like that ... D. __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com |
From: Tommy H. <th...@ch...> - 2003-01-29 07:00:43
|
Hi, To those who don't know me, I'm currently an employee of KnowNow and I've been working on some Windows connectivity stuff on my own time. I'm getting close to releasing a version of my stuff that is based on a C++ framework that has the following feature: * Supports all Windows platforms: Win98, WinME, WinNT, Win2K, WinXP, and WinCE 3.0. * Multiple interfaces to the C++ code: COM/ActiveX and .NET. * Thread-safe supporting multithreaded applications. * Doxygen comments. * cppunit-based unit tests. * Sample applications. * Based on WinInet to support proxies and SSL easily. I don't know the exact timeframe of when I will be releasing it, but it is soon. However, I don't expect people to wait until I release it. If you are working on something native to .NET, there's no reason why there can't be more than one version. I would be interested in other versions of any client libraries. Either Adam or I will make an announcement when it becomes available. Thanks. Tommy -----Original Message----- From: Asynch Messaging [mailto:asy...@ho...]=20 Sent: Thursday, January 23, 2003 11:15 AM To: mod...@li... Subject: Re: [Mod-pubsub-developer] libkn released; notes from 1/2003 MPSUG > ----- Original Message ----- > From: "Adam Rifkin" <Ad...@Kn...> > > ACTION ITEM #2: Once this tarball ships, the next step is to "Windows-enable" c_pubsub so we have connectivity to > Windows desktops, as Mike Dierken and Jeff Barr have discussed in the=20 > past and as KnowNow currently offers. This sounds cool. What's the thinking? Are we talking about a COM object that raises events that are visible from scripted clients - essentially a COM interface to a 'router' object with subscribe()/publish()/unsubscribe() methods? If so, I've got some preliminary work in that area - I just need to originating event to toss up. > 2. We acknowledged how much we love Mike's AmazonPop application,=20 > but noted that it crashes > ungracefully "in stupid, ugly, unpredictable ways". And unreported ways - isn't there a bug system on SourceForge? Not that I'd actually do anything about the problems, but at least they'd be listed somewheres... > ACTION ITEM: Think about a mod_pubsub weblog. Can any good come from having one, or should we just > stick to the mailing list for now? I think a persistent log of thoughts and surface discussion is useful, especially if it ties in other technology and ideas. It's often easier than a mailing list archive. It also gets into blogs which are growing in their use of pub/sub concepts. So I vote yes on a weblog. ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld =3D Something 2 See! http://www.vasoftware.com _______________________________________________ Mod-pubsub-developer mailing list Mod...@li... https://lists.sourceforge.net/lists/listinfo/mod-pubsub-developer |
From: <kr...@po...> - 2003-01-27 20:03:27
|
Derek Robinson writes: > I wrote a proof-of-concept demo of 'Pushlets' in PHP a few months ago for > my own amusement. I concluded (as did the Java Pushlets guy) that pushing > content over http to more than 20 or so clients puts an unnecessarily big > burden on the server. (In fact the Pushlet guy has been pushing 'pullets' > for quite a while now.) Well, it depends. Most web servers can handle ten hits a second without any trouble; assuming 10 kilobytes per hit, 20 hits a second saturates a T1. If you have no inline images, and your users spend an average of 60 seconds on each page, that's only 1200 users. And if you're trying to get out real-time updates to these folks by means of "pull", you have a maximal latency of 60 seconds; if you want updates to be twice as frequent, it takes twice the bandwidth and server load. Now, by contrast, IRC servers on the Undernet regularly handle 20 000 clients. When I first performance-tested the push_manager included in mod_pubsub back in 2000, it could handle 4 000 simultaneous push connections on a 1GHz box without any trouble, and that's on Linux without any compile-time kernel tuning. IRC servers generally run on FreeBSD because its network stack scales better, or at least used to. Typically, an IRC client sends a notification every 10-20 seconds, and receives a message 1-10 times a second. So, anyway, pushing out updates when they happen can impose a much smaller load on the server's bandwidth than having the client poll in common circumstances. > But hey, if anyone's interested I could dig it out & dust it off ... I'd love to see it. -- <kr...@po...> Kragen Sitaker <http://www.pobox.com/~kragen/> Edsger Wybe Dijkstra died in August of 2002. The world has lost a great man. See http://advogato.org/person/raph/diary.html?start=252 and http://www.kode-fu.com/geek/2002_08_04_archive.shtml for details. |
From: <kr...@po...> - 2003-01-27 20:03:27
|
Adam asks: > 7. We're curious if Kragen is currently brewing anything in the > background. Only in my own mind, I fear. I'll let you know. > 9. Monthly MPSUG meetings are fun socially and let us chat about = > what's going on with the package. > > ACTION ITEM: I suggest we regularly meet somewhere in Silicon Valley = > on the last Tuesday of the month at 8pm (this month excepted). We'll = > pick a different restaurant each time, announce it on the "developer" = > list, and keep it open to anyone who wants to come and discuss. I hope we can have these someplace north of Palo Alto, because I'd like to come. By the way, the last Tuesday of this month is tomorrow, and Ben and I already plan to be working on mod_pubsub at my house at 8PM anyway (although I have to go to bed not long after that); if other people want to come, maybe we could get even more done. -- <kr...@po...> Kragen Sitaker <http://www.pobox.com/~kragen/> Edsger Wybe Dijkstra died in August of 2002. The world has lost a great man. See http://advogato.org/person/raph/diary.html?start=252 and http://www.kode-fu.com/geek/2002_08_04_archive.shtml for details. |
From: Scott A. L. <sc...@sc...> - 2003-01-27 17:25:56
|
Interesting related thread on converging IM with weblog software. (View the comments for the discussion.) <http://radio.weblogs.com/0001015/2003/01/26.html#a1691> scottandrew |
From: Asynch M. <asy...@ho...> - 2003-01-27 17:23:44
|
----- Original Message ----- > > Actually, I think I see your point now. Instead of having to ping > several dozen subscribers, Why do they call it 'ping'? I always associate 'ping' with checking to see if the target exists - like the 'ping' command line utility in Unix and Windows. |
From: Scott A. L. <sc...@sc...> - 2003-01-27 07:15:45
|
Scott Andrew LePera wrote: > Very much agreed. You wouldn't believe the crap I went through to get to > that weblogs.com thingy working, and that's only one or two pieces of > info. The weblogs.com format sends the least amount of info, but > requires more work to process. TrackBack is just the opposite: more > info, less work. Oops, I should mention that weblogs.com itself DOES support a plain vanilla POST interface [1], but Dave told me himself he prefers XML-RPC [2]. Hehe. scottandrew [1] <http://newhome.weblogs.com/directory/11/howToPing> [2] <http://scriptingnews.userland.com/backissues/2002/04/18#l66952626e8ae873cdf07c75235028086> |
From: Scott A. L. <sc...@sc...> - 2003-01-27 07:06:18
|
Asynch Messaging wrote: > Notifications are notifications, and isn't less useful because they are > really fast. Yes, but they aren't necessarily /more/ useful, yet, which is why people haven't abandoned things like email notifications and RSS polling. Polling every 5 minutes is "good enough" for most; otherwise, I figure someone would've hooked up Jabber for real-time notification delivery awhile ago. > Did you know that mod_pubsub will do off-site posting via HTTP POST? > This means that mod_pubsub will relay the notification (aka 'ping') to > multiple blogs, not just live in-browser clients. You can already notify multiple blogs without url routing. "Live" reactive endpoints (not just in-browser) are part of what makes it compelling. But yeah, that's only part of the mod_pubsub story. > What is wrong with per-entry pub/sub? Nothing! > There should be other means to subscribe to the blog itself. > Think of the blog as a 'collection of entries' and subscribe to that > collection. Whenever something is added (a POST), you get notified. Yeah, this is originally what I proposed to Ben when they were tinkering with the TrackBack idea. They took it deeper, applying it to post-to-post relationships as well as post-to-category/directory. Actually, I think I see your point now. Instead of having to ping several dozen subscribers, interested parties can subscribe to one URL and offload the work onto the router. So, yes, it would be cool if weblog software were able to emit a non-contextual, TrackBack-compatible notification with every update. >>Personally, I'd like to see the TrackBack ping supplant the weblogs.com >>ping, because 1) it has more information, and 2) doesn't require XML-RPC. > > Me too. It'll happen over time. RPC helps some quickly bang out demos, but I > like REST for the long term. Very much agreed. You wouldn't believe the crap I went through to get to that weblogs.com thingy working, and that's only one or two pieces of info. The weblogs.com format sends the least amount of info, but requires more work to process. TrackBack is just the opposite: more info, less work. scottandrew |
From: Asynch M. <asy...@ho...> - 2003-01-27 02:00:12
|
----- Original Message ----- From: "Scott Andrew LePera" <sc...@sc...> > Maybe because real-time notifications in blogland have dubious value, at > least right now? ;) Notifications are notifications, and isn't less useful because they are really fast. > > I actually put up a mod_pubsub app that accepts the weblogs.com ping > from weblog software that supports it: > > http://mod-pubsub.thrillerguide.com/blogtalk/ > > Send a weblogs.com ping to: > http://mod-pubsub.thrillerguide.com/blogtalk/blogtalk.php > > ...and your blog name will appear in the textarea. Extremely boring. > Much like a real-time weblogs.com listing. The problem is, the ping > doesn't send much interesting info. Did you know that mod_pubsub will do off-site posting via HTTP POST? This means that mod_pubsub will relay the notification (aka 'ping') to multiple blogs, not just live in-browser clients. So, what you have with mod_pubsub is more than live "in-browser" clients, you have url-to-url routing. > > Now, TrackBack/PingBack is /much/ more interesting. However, TrackBacks > are usually associated with a specific post, or at most, a Movable Type > category group. To my knowledge, most weblog software doesn't send a > TrackBack ping when a blog is updated "in general." What is wrong with per-entry pub/sub? There should be other means to subscribe to the blog itself. Think of the blog as a 'collection of entries' and subscribe to that collection. Whenever something is added (a POST), you get notified. > > Personally, I'd like to see the TrackBack ping supplant the weblogs.com > ping, because 1) it has more information, and 2) doesn't require XML-RPC. Me too. It'll happen over time. RPC helps some quickly bang out demos, but I like REST for the long term. |
From: Scott A. L. <sc...@sc...> - 2003-01-27 00:26:31
|
Asynch Messaging wrote: > You know, theres a lot of web native pub/sub going on & I don't see > mod_pubsub anywhere on the scene. Maybe because real-time notifications in blogland have dubious value, at least right now? ;) I actually put up a mod_pubsub app that accepts the weblogs.com ping from weblog software that supports it: http://mod-pubsub.thrillerguide.com/blogtalk/ Send a weblogs.com ping to: http://mod-pubsub.thrillerguide.com/blogtalk/blogtalk.php ...and your blog name will appear in the textarea. Extremely boring. Much like a real-time weblogs.com listing. The problem is, the ping doesn't send much interesting info. Now, TrackBack/PingBack is /much/ more interesting. However, TrackBacks are usually associated with a specific post, or at most, a Movable Type category group. To my knowledge, most weblog software doesn't send a TrackBack ping when a blog is updated "in general." Personally, I'd like to see the TrackBack ping supplant the weblogs.com ping, because 1) it has more information, and 2) doesn't require XML-RPC. scottandrew > Blogs have 'TrackBack' and 'PingBack' (an xml-rpc variant) for web > notifications. The blog community uses 'ping' where I would have used > 'notify', but what the heck. > > I would like to see how mod_pubsub can be used with TrackBack. The format > for TrackBack is POST with "application/x-form-www-urlencoded", using > specific named values, so it isn't complicated. > > > == > http://www.popdex.com/blog/archives/000029.html > Popdex now offers the ability to host TrackBack threads for any arbitrary > URL > TrackBack pings (notifications) can now be sent to Popdex for any arbitrary > URL, thus creating a hosted, third-party TrackBack solution for any URL on > the web. > > My idea is that maybe there will at some point be several "Trackback > Aggregation Services", in a similar way that there are several blog update > monitoring services, mainly weblogs.com and blo.gs. > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: > SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! > http://www.vasoftware.com > _______________________________________________ > Mod-pubsub-developer mailing list > Mod...@li... > https://lists.sourceforge.net/lists/listinfo/mod-pubsub-developer > > |
From: Asynch M. <asy...@ho...> - 2003-01-26 23:32:18
|
You know, theres a lot of web native pub/sub going on & I don't see mod_pubsub anywhere on the scene. Blogs have 'TrackBack' and 'PingBack' (an xml-rpc variant) for web notifications. The blog community uses 'ping' where I would have used 'notify', but what the heck. I would like to see how mod_pubsub can be used with TrackBack. The format for TrackBack is POST with "application/x-form-www-urlencoded", using specific named values, so it isn't complicated. == http://www.popdex.com/blog/archives/000029.html Popdex now offers the ability to host TrackBack threads for any arbitrary URL TrackBack pings (notifications) can now be sent to Popdex for any arbitrary URL, thus creating a hosted, third-party TrackBack solution for any URL on the web. My idea is that maybe there will at some point be several "Trackback Aggregation Services", in a similar way that there are several blog update monitoring services, mainly weblogs.com and blo.gs. |
From: Asynch M. <asy...@ho...> - 2003-01-24 07:39:08
|
How about a C# client? Mono looks to be fairly decent... http://www.go-mono.com/ |
From: Derek R. <nn...@ya...> - 2003-01-23 23:36:28
|
> 6. We are still on a crusade to PHP-enable mod_pubsub. It's unclear > to some folks what PHP on the server side buys them, since PHP is > generally used as "better than ASP or JSP for serving static web pages" > and inherently uses a request/response model. Perhaps a PHP "client" > (a/k/a "microserver") of mod_pubsub is a better use of time than porting > the Perl mod_pubsub cgi scripts. I wrote a proof-of-concept demo of 'Pushlets' in PHP a few months ago for my own amusement. I concluded (as did the Java Pushlets guy) that pushing content over http to more than 20 or so clients puts an unnecessarily big burden on the server. (In fact the Pushlet guy has been pushing 'pullets' for quite a while now.) But hey, if anyone's interested I could dig it out & dust it off ... __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com |
From: Joyce P. <tru...@ya...> - 2003-01-23 21:30:20
|
Since we're all sharing, here are the notes from the first MPSUG meeting in December. Meeting of the mod_pubsub user's group December 11, 2002 Attending: 0. Scott LePera 1. Joyce Park 2. Adam Rifkin 3. Ben Sittler 1. Joyce announced that immediately before the meeting, she had packaged and released version 0.3 with all contributed code to date. She also reiterated that Ben also has permission to package and release code. Ben suggested that he might take over some of the release tasks of the group. 2. Adam enumerated his efforts to stir up support among influential developers. The most important single event was Rohit Khare's announcement of mod_pubsub at the Supernova conference, which led to mentions on ScriptingNews.com and other high-traffic weblogs. Adam has also written to numerous influential Web developers, many of them former employees of KnowNow, Inc., with invitations to try the product and perhaps join the developer community. Response has been gratifyingly positive given the generally poor relationship between KnowNow, Inc. and said developers -- with perhaps a whiff of feeling that mod_pubsub was "old news". Mike Dierken (AsynchMessaging) has contributed actual code to the project. 3. Scott, as Head Webdev, was asked to add a link to the Sourceforge project page from the homepage. This necessitates him becoming more comfortable with the Sourceforge file release system, which is idiosyncratic. 4. Joyce, as fit punishment for her constant bitching about the slowness and ugliness of the Perl module, has been tasked with porting to PHP. Alternatively or in the meantime, Ben has been tasked with looking into porting via Swig. The PHP community has been totally ignored by KnowNow, Inc. to date, a significant omission given the size of the community and its affinity for our product. We also discussed our technical goal, of producing a full Apache module written in C, and why the easiest way to get there might be via a PHP port. PHP already has excellent HTTP handling functionality built in C, as opposed to the clutch of Perl libraries used by the Perl module. 5. Ben was tasked with checking in libkn, the so-called "C microserver", which Adam informed us will be contributed by KnowNow, Inc. 6. Joyce suggested that development of applications be forked to reflect the philosophical disagreement between mod_pubsub developers who wish to support all browsers and those who wish to take advantage of the features of recent browsers. All agreed that this should be done, with Ben's proviso that if he could find a way to make non-NS4 apps compliant with NS4, he would do so on the NS4 branch. 7. Many expressions of wonder and thanksgiving were directed at the current management of KnowNow, Inc. for their magnanimity in releasing mod_pubsub's codebase. All attendees mentioned that work would likely slow down due to upcoming holidays and vacations, and Joyce and Ben's lack of connectivity for several weeks. __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com |
From: Asynch M. <asy...@ho...> - 2003-01-23 19:13:36
|
> ----- Original Message ----- > From: "Adam Rifkin" <Ad...@Kn...> > > ACTION ITEM #2: Once this tarball ships, the next step is to "Windows-enable" c_pubsub so we have connectivity to > Windows desktops, as Mike Dierken and Jeff Barr have discussed in the past and as KnowNow currently offers. This sounds cool. What's the thinking? Are we talking about a COM object that raises events that are visible from scripted clients - essentially a COM interface to a 'router' object with subscribe()/publish()/unsubscribe() methods? If so, I've got some preliminary work in that area - I just need to originating event to toss up. > 2. We acknowledged how much we love Mike's AmazonPop application, but noted that it crashes > ungracefully "in stupid, ugly, unpredictable ways". And unreported ways - isn't there a bug system on SourceForge? Not that I'd actually do anything about the problems, but at least they'd be listed somewheres... > ACTION ITEM: Think about a mod_pubsub weblog. Can any good come from having one, or should we just > stick to the mailing list for now? I think a persistent log of thoughts and surface discussion is useful, especially if it ties in other technology and ideas. It's often easier than a mailing list archive. It also gets into blogs which are growing in their use of pub/sub concepts. So I vote yes on a weblog. |
From: Adam R. <Ad...@Kn...> - 2003-01-23 18:57:58
|
Hi mod_pubsubbers! Long time, no hear... it's been almost a month since = our last post. (Talk about a low-traffic list!) So I'm taking time off = from my vacation to post an update. On Tuesday, January 21, 2003 we had a Mod PubSub User Group (MPSUG) = meeting with five attendees who choose to remain nameless. Here are the = minutes: 1. Fred Sanchez's C library that connects to mod_pubsub, which used = to go by the name libkn, is now a part of the mod_pubsub package, = checked into the CVS directory c_pubsub. ACTION ITEM #1: Joyce Park will put a new "ginormous" tarball on = Sourceforge that includes this component. ACTION ITEM #2: Once this tarball ships, the next step is to = "Windows-enable" c_pubsub so we have connectivity to Windows desktops, = as Mike Dierken and Jeff Barr have discussed in the past and as KnowNow = currently offers. ACTION ITEM #3: KnowNow can then use c_pubsub at KnowNow to produce a = "Microserver SDK" package that can be used both with KnowNow's = commercial product and with mod_pubsub. This gets KnowNow some bang for = the buck from the mod_pubsub project, and gets mod_pubsub some = credibility from commercial use. 2. We acknowledged how much we love Mike's AmazonPop application, but = noted that it crashes ungracefully "in stupid, ugly, unpredictable = ways". One thing we'd love to see is a "best practices" paper -- like = Scott LePera's "Busy Developer's Guide to mod_pubsub" idea, that tells = people "how to make shit that doesn't suck ass". (Forgive the phrasing = but hey, the shoe fits. :) The paper should be a short document on = graceful failure & how to do it in the least effort-requiring, most = stylistic way. ACTION ITEM: Scott, Ben Sittler, and I will sit down to outline the = "best practices" paper, and we'll run it by anyone (Mike? Joyce?) who's = interested in giving feedback and has time to do so. 3. Joyce graciously purchased the modpubsub.com and modpubsub.org = domain names. These do not belong to KnowNow; they belong to the = mod_pubsub community. ACTION ITEM: Joyce will make the home page at Sourceforge forward to = modpubsub.org ...=20 4. ... but we could still use a mascot/logo. Suggestions included a = tavern sign (Pub) and sandwich (Sub), or a PUG (for Pubsub User Group). = No decisions were made. ACTION ITEM: Think about a logo/mascot. ACTION ITEM: Think about a mod_pubsub weblog. Can any good come from = having one, or should we just stick to the mailing list for now? 5. We talked some about improving the Protocol document included with = the distribution. ACTION ITEM: Adam and Ben will investigate this. This is probably = useful to KnowNow as well. 6. We are still on a crusade to PHP-enable mod_pubsub. It's unclear = to some folks what PHP on the server side buys them, since PHP is = generally used as "better than ASP or JSP for serving static web pages" = and inherently uses a request/response model. Perhaps a PHP "client" = (a/k/a "microserver") of mod_pubsub is a better use of time than porting = the Perl mod_pubsub cgi scripts. ACTION ITEM: Discuss. 7. We're curious if Kragen is currently brewing anything in the = background. ACTION ITEM: Ben will get in touch with Kragen. =20 8. There's still no index of the 30something applications that ship = with the mod_pubsub package. ACTION ITEM: Adam will make an index page in February and check it = into CVS. This is probably useful to KnowNow as well; they have already = shown interest in using some of these applications as samples included = in KnowNow's forthcoming "Live Browser" commercial offering. 9. Monthly MPSUG meetings are fun socially and let us chat about = what's going on with the package. ACTION ITEM: I suggest we regularly meet somewhere in Silicon Valley = on the last Tuesday of the month at 8pm (this month excepted). We'll = pick a different restaurant each time, announce it on the "developer" = list, and keep it open to anyone who wants to come and discuss. That's all for now. More soon... Adam http://mod-pubsub.sourceforge.net/ mod_pubsub statistics as of January 23, 2003 # certified as mod_pubsub developers: 36 # on mailing lists: 28 on developer, 24 on general # of mailing list posts: 59 CVS Repository: 7 commits, 146 adds # of downloads: 101 # of pageviews: 3412 overall ranking: 7141 out of 55,112 (29.64 percentile) |
From: <kr...@po...> - 2002-12-26 11:42:49
|
Joyce Park wrote: > [quoting Kragen] > > I wrote clENS (now pubsub.py) to make it easier --- if > > you have Python installed, you just run the script, and it works. > Are Windows users any more likely to have Python than Perl? No, but Microsoft Windows users can more easily install Python than Perl, URI, Digest::MD5, MIME::Base64, LWP, Apache, mod_perl, and probably some other Perl module I forgot. I never did care enough to package up pubsub.py with the McMillan Installer so you didn't even have to have Python installed, but if that matters a lot to someone, it should only take an hour or two. > How about if I rewrite the installation doc to have httpd.conf as the > default, but also give directions for the .htaccess thing? One of my > issues is that I found I had to set AllowOverride to All for the > default, docroot, AND mod_pubsub/cgi-bin directories to get this to > work. Very annoying, not the way it was documented, insecure, and > you wouldn't be able to do that on most shared servers anyway. Let's we just rip out the .htaccess bit; the justifications that were dubious in 2000 don't apply at all now. A savvy user can always put the httpd.conf text into .htaccess and have it work fine, anyway, in the unlikely case that they want to go with .htaccess. Maybe you already did this; I haven't updated from CVS in a while, and as I write this, I have no Internet access. -- <kr...@po...> Kragen Sitaker <http://www.pobox.com/~kragen/> Edsger Wybe Dijkstra died in August of 2002. The world has lost a great man. See http://advogato.org/person/raph/diary.html?start=252 and http://www.kode-fu.com/geek/2002_08_04_archive.shtml for details. |