Alek Links Etc.

Java-based New Open-Source BPEL Engine Release to be Generally Available This Month @ SYS-CON INDIA:

(...) Active BPEL, LLC announced the latest milestone release of its open-source BPEL engine, which includes several feature enhancements, as well as updates on its planned WS-* support. The ActiveBPEL engine is an Open Source implementation of a BPEL engine, written in Java. ActiveBPEL was released into open source in July 2004 by Active Endpoints, Inc. In late October, Active Endpoints released metrics for ActiveBPEL, stating that the engine had achieved more than 20,000 "quality" downloads (attributed directly to an operating organization or a recognized ISP), and had more than 400 registered community participants. Most of the downloads and implementations have been in the technology, government, and telecommunications sectors, the vendor reported. Milestone 2 adds direct invocation of another BPEL process running on the same engine; ability to set variables; enhanced process query capabilities; XML database support Software AG's Tamino product, and Tomcat 5.5 and Java 1.5 support. The product also includes support for WS-* specifications, including WS-Addressing and WS-Policy. However, a release from the organization states that support for WS-Security and WS-ReliableMessaging, originally planned for version 2.0, have been deferred until a future release. ActiveBPEL recommends that developers interested in adding these specifications "use custom invocation handlers or use external Web services that implement those specifications as interfaces." ActiveBPEL also stated that an upcoming version 3.0 would be announced in the first quarter of this year, and will be based on the upcoming WSBPEL 2.0 specification, while continuing to support processes that were developed using the BPEL4WS 1.1 specification(...)

MagicDraw UML 10.5 Extends Business Process Modeling with Export to BPEL - PR.com:

(...) The new release of MagicDraw also delivers brand new integration with ProActivity Business Process Analysis Suite, BPEL export to BEA WebLogic 8.1, updated integration with NetBeans and improved GUI.
Allen, TX, December 30, 2005 --(PR.COM)-- No Magic, Inc.®, a leading vendor of architecture modeling software, today announced the release of MagicDraw® UML 10.5, an upgrade of No Magic?s award-winning UML-based architecture tool. MagicDraw 10.5 extends its business process modeling capabilities to export Business Process Execution Language (BPEL 1.1) compliant code from Business Process Modeling Notation (BPMN) diagram. The new version delivers a brand new integration with ProActivity Business Process Analysis suite, updated NetBeans integration and improved GUI to increase flexibility and convenience using the tool.
No Magic continues its commitment to narrow the divide between business needs and the systems developed to meet those needs by extending a Business Process Modeling offering in its award-winning UML-based architecture tool. The new MagicDraw UML BPEL Export functionality allows the users to export their Business Process Execution Language (BPEL 1.1) compliant code from a Business Process Modeling Notation (BPMN) diagram. The outputted BPEL code can be used in BEA WebLogic 8.1. (...)

HBO Forums: Loca in America (posted by LemniscusTuborum)

(...) In Florida: Vicus Filii Gaii (=Jacksonville)
Oppidum Lavantis (= Washington)
Provincia Porci Novi (= New Hamshire)
Remultum = Muchagain (where Detroit is)
Tunica Puellae Super (= Gal-Vest-On, Texas) (...)

What an interesting idea! ROME in Latin Subtitles Petition Thread (HBO Forums). Alas, it seems that the first ROME DVD is already in post-production nonetheless somebody already translated to Latin few episodes as a DEMO!

(...) Wouldn't it be too cool if you could later on buy a DVD set of the ROME series with among the subtitles the Latin language to choose for? Or even among the language settings?? If there is anyone out there who wants this or is interested, to have yet another great detail added to a great series, let this thread know!(...)

Song used in "Frantic"? [Archive] - Harrison Ford Forum:

(...) does anyone know the title and artist of the song playing when Dr. Walker enters the bar "Blue Parrot" for the first time? I don't know how to find out.... (...)

BPEL 2 gets delayed (News & Analysis - Network Magazine India:

(...) Version 2.0 of the Web services business process execution language (BPEL) will have to again wait till next year. The Organisation for the Advancement of Structured Information Standards (OASIS) is yet to give the go-ahead to enable human interactions as a part of this technology.
BPEL was designed to co-ordinate between Web services considered critical for applications such as transactions and B2B systems. With increasing portal initiatives among enterprises, the technology has wide industry support from vendors like Microsoft, BEA and IBM. The preliminary BPEL 1.1 and 2.0 technologies have been incorporated by Oracle and WebSphere in Oracle BPEL Process Manager and IBM Websphere Process Server respectively. However, OASIS has to accept it as an official specification.
In addition to BPEL, OASIS coordinated the design of BPEL with that of BPELJ (Java) and the Web services remote protocol (WSRP). Due to the standardised Web services supported by Java, XML and other platforms, it has created scope for the introduction of third-generation portals.
The first versions of this new generation of portals were released in 2004, but by 2006-end we should see considerably more innovative inter-operability and BPM tools as OASIS protocols such as WSRP are expanded and extended.
BPEL has functions such as language constructs to enable if-then-else statements for the 2.0 version. According to IBM, dynamic, parallel invocation of services is also an important addition for adapting the number of steps in a process based on the number of partners participating in a transaction.
Experts are of the view that version 2 improves the way Web services are used to call out to partner links to enable more flexibility. However, issues such as merging and creating documents still need to be addressed. (...)

Delightfully self referrential (or recursive) ... Modesty Blaise by Peter O'Donnell ( more details and the cover made just for Pulp Fiction can be seen in Wikipedia) This amazon review gets it short: Seen "Pulp Fiction"? Here's One of the Stars of the Movie, August 31, 2005

(...) Reviewer: Chris Ward (Costa Rica) (...)
Okay, so you've seen Quentin Tarantino's movie "Pulp Fiction." And you've noticed that every time Vincent Vega (John Travolta) goes to the bathroom, he takes a book along to read. What's he reading? "Modesty Blaise."
Why would Vincent want to read Modesty Blaise? Because he's cool and Modesty's cool-- she's a product of British "cool," circa 1965-- and it doesn't get any cooler than that. Meet Modesty and her pal Willie and join them on their action-packed adventures, and you'll see why Tarantino worships Modesty, and why he (pretty obviously) patterned much of Uma Thurman's character "The Bride" after Modesty in the movie "Kill Bill." Modesty Blaise is the ultimate action heroine-- try this, her first adventure (in book form, anyway-- she'd been a comic book figure for awhile by the time this came out), and enjoy. (...)

Applies ot e750 as well (it just has no software i never had use for anyway ...):

Good: reading books is pleasure (with uBook), decent music player, so-so note taker

Bad: pressing record button is too easy to happen accidentely when the PDA in a pocket - fortunately this button canbe configured to do nothing!

Ugly: runs out of memory, manual killing helps but eventually reboot is required after few days/weeks of usage to clean up garbage; UI issues with controlling battery saving and making PDA do what you want

Follow white rabbit ... ahem newsgroup thread via Krzysztof Kowalczyk weblog: Perl and lisp programmers

(...) There's no reason perl programs have to be unreadable any more than there is a reason that Lisp programmers have to emit a continuous whining sound. (...)

big names were mentioned ... ahem heavy artillery was used:

(...) People do cool stuff with the most primitive tools you can imagine (and Java is not the most primitive tool). Beethoven wrote fantastic music without the aid of electricity and people play it on things made out of wood, bone glue and animal guts, some of them hundreds of years old. Some of the best things he wrote *after he'd gone deaf*! And if you want cool tools, Java has them. Look at eclipse for instance: a pretty comprehensive IDE with loads of cool stuff in it. Look at aspectj. Yes, I know, Lisp has, or could have, or once had, better versions of all these things: so what? Most of all: just use the tools you have to do interesting stuff, instead of sitting around whining. (...)

and some subversive techniques to get around "java issue":

(...) A friend of mine works in a java shop and he uses Jython, writes all his code with python syntax and delivers jars. No one has caught on that he is programming in python, since he produces more than anyone elso, no one cares. (...)

and other observations

(...) > Real hackers program directly in hex code. Assembly is a girl thing.
Once you get beyond an oscilloscope, its all gui crap. (...)

UNIX Socket FAQ - blocked by sockets in time-wait??:

(...)RobSeace (...) Yes, it sounds quite likely that it's some sort of anti-DOS thing... Do you see any logged kernel messages in "dmesg" or anything? Have you played around with the various "/proc/sys/net/ipv4/" settings to see if it improved things at all?
I really hate the SO_LINGER approach, because avoiding the TIME_WAIT state is dangerous... But, if it works with a non-zero linger time (ie: it seems to recover from its stuck state after that given linger time elapses), then I think the problem is not so much TIME_WAIT as more likely FIN_WAIT*... Because, if the remote FIN arrived within the given linger time, everything should proceed as normal, and the socket would still go TIME_WAIT; but, if it failed to arrive, it'd sit in FIN_WAIT_1 or FIN_WAIT_2 until the linger time expired, at which point, it'd RST the connection and just fully tear down the socket... That indicates to me that it is indeed the Windoze client host that's getting wacked out somehow, and starting to drop/ignore FINs... (Does "netstat" show FIN_WAIT_* sockets on the server system?) I don't have nearly as much problem with this sort of use of SO_LINGER as I do with using with a 0 linger time, to just always RST connections, and always avoid the TIME_WAIT state... At least in this case, most normal connections should still get a TIME_WAIT, but badly behaved ones that can't be shutdown properly in a reasonable amount of time get RST... But, I still just dislike SO_LINGER on general principle, anyway... ;-)
You might want to try setting up a sniffer and see if you can spot the change in behavior that triggers this when it starts happening... If it's really some detectable change in behavior anyway, and not just the Linux box doing some anti-DoS trickery on the lone host it seems pounding on the same port over and over again... *shrug* (I'd think that if it were some sort of anti-DoS thing, it would at least log something somewhere about it... And, probably be tunable, somehow... Hmmmm... You're not running any kind of netfilter rules on the server host are you?) Reply With Quote (...)

NBIO: Java Non-blocking I/O Library:

(...) If you are running on Linux 2.2.x systems you may wish to increase your file descriptor limit, which increases the number of simultaneous socket connections the system can have. This is important as one of the basic uses for NBIO is to write server applications which can support many simultaneous connections (many more connections than were possible using threads). This is relatively simple to do:
Create the file /etc/initscript (if it does not already exist) and place the following commands into it:

umask 022
ulimit -c 2097151
ulimit -Hn 32768
PATH=/bin:/sbin:/usr/bin:/usr/sbin
export PATH
eval exec "$4"

If /etc/initscript already exists, add the command

ulimit -Hn 32768

to it.
Add the following commands to /etc/rc.d/rc.local (or some other startup script which is run at boot time):

echo 32768 > /proc/sys/fs/file-max
echo 65536 > /proc/sys/fs/inode-max

Add the following command to your login scripts (e.g., .bashrc or .cshrc):

ulimit -n 32768

Reboot the system for the above changes to take effect. (...)

From lide 9 in lecture notes to TCP/IP in G22.3250 -- Honors Operating Systems (via Ken):

(...) TIME WAIT Problems with closed socket
- What if final ack is lost in the network?
- What if the same port pair is immediately reused for a new connection? (Old packets might still be foating around.)
Solution: "active" closer goes into TIME WAIT
- Active close is sending FIN before receiving one
- After receiving ACK and FIN, keep socket around for 2MSL (twice the .maximum segment lifetime.)
Can pose problems with servers
- OS has too many sockets in TIME WAIT, slows things down
- Hack: Can send RST and delete socket, set SO LINGER socket option to time 0 (useful for benchmark programs) (...)

Amazing how a little typo can give that much grievence ... there is some way to chnage it by modifying Windows Registry (and rebooting ...).

From UNIX Socket FAQ - blocked by sockets in time-wait??:

(...)RobSeace Location: Boston, MA (...)
(...) Yeah, with ~4000 TIME_WAIT sockets, you're covering the entire range of 1024-5000, which I'm betting anything the Windoze host is using as its ephemeral port# range (and, lots of other systems for years and years have stupidly used, due to a silly typo in early BSD)... So, you'd have a TIME_WAIT socket for every possible ephemeral port the Windoze box might choose, so when it attempts to connect with one, it'd get treated as a wandering duplicate packet from the previous connection that used that port#... At that point, the Windoze box probably tries to assassinate the TIME_WAIT socket on the Linux box, but I believe Linux protects against TIME_WAIT assassination, and so would probably ignore the RSTs from the Windoze box, and leave its TIME_WAIT intact for the duration of the timeout... But, this is all just speculation on my part; if you set up a sniffer and looked at the actual traffic that occurs during this weirdness, you'd be able to tell for sure...(...)

ONJava.com: The PHP Scalability Myth:

(...) # Yes lets compare apples and oranges...
2003-10-17 11:34:34 anonymous2 [Reply | View]
Let's remove PHP/JAVA and replace with Apples/Oranges...
Article Summary: "Apples are better since they have a red peel. And if under ripe they can be orangish in color. Thus they are superior then oranges, since oranges can't be red!"
I am astounded that this article was even published perhaps it would have gone better on a blog or a piece of toliet paper, (on the floor of a rest stop)...
(...) # Yes lets compare apples and oranges... 2003-10-17 17:13:49 anonymous2 [Reply | View]
Blood oranges are red. Oranges are really superior to apples in every way.

* I like apples and oranges. 2003-10-27 05:57:25 anonymous2 [Reply | View]
So your saying that specific sorts of oranges are better then apples?
Oh, and apples are easier to eat then oranges, because you don't have to peel them. Apples are a bit less messy then oranges too. You have to be really careful that you don't spray juice everywhere with oranges, and even so you almost always end up hitting someone in the eye.
Apples come in green too! Haven't seen many orange apples to be honest.

o I like apples and oranges. 2003-11-19 18:22:03 anonymous2 [Reply | View]
"You have to be really careful that you don't spray juice everywhere with oranges, and even so you almost always end up hitting someone in the eye."
Then use a different method.
If an orange isn't orange, is that bad documentation or a poor implementation? (...)

Just watched on BookTV an interview with William Sampson author of Confessions of an Innocent Man : Torture and Survival In a Saudi Prison (Amazon). He made a point that out of 6.5B people only 1B are in countries that provide reasonable level of protection against unlawful detainment and torture and that to maintain moral highground one can not make any compromises - there is no 'torture light' ...

(...) On Sunday, December 17, 2000, Canadian engineer William Sampson stepped outside his house in Riyadh only to be hauled into a car and beaten by two Saudi men he didn't know. Within an hour, he was incarcerated in one of the city's most notorious jails. Within two months, he was tortured into a confession of responsibility for a wave of car bombings he did not commit. Sometime in that first year, he was sentenced to death in a secret trial. For two and a half years, Sampson was continually subjected to beatings and torture, convinced his death was just around the corner. Inept diplomacy failed him but human rights groups took up his cause and on August 8, 2003, he was finally freed in a controversial prisoner exchange. It wasn't until February 2005 that Sampson's name was officially cleared when a British inquest exonerated him of the crimes.
Angry, intelligent, and compelling, Sampson places his personal story within the context of the geopolitics that engineered his fate, and in doing so has crafted a searing exposé of Western foreign policy in the Arab Middle East.
About the Author William Sampson was working as a marketing consultant in Riyadh at the time of his arrest. He holds a Ph.D. in biochemistry and an MBA from Edinburgh University. The British Court of Appeal recently awarded Sampson and his fellow detainees the landmark right to sue their torturers in Saudi Arabia. (...)

I noticed that when Kerio Firewall 4 warned that acdsee 2.42 when finishing tries to launch netsonic - scary ... as mentioned in review of ACDSee 32 v.2.42 (no longer available)

(...) There were few annoyances to be found in ACDSee 32. For an unknown reason, an advertisement popped up in its own window for a software application called "Net Sonic" upon closing ACDSee 32. Fortunately the window contained a "Remind me later" check box that I unchecked. An odd addition to a commercial (registered) application.

The really good part about jvmstat is that it is designed to not affect JVM performance when running and that is much different from profiler ... :

(...) The jvmstat technology adds light weight performance and configuration instrumentation to the HotSpot JVM and provides a set of monitoring APIs and tools for monitoring the performance of the HotSpot JVM in production environments. The instrumentation is designed such that it is 'always on', yet has negligible performance impact. The monitoring interfaces added to the HotSpot JVM are proprietary and may or may not be supported in future versions of the HotSpot JVM.
(...) As of J2SE 5.0 and later, a subset of the jvmstat tools is included with the JDK (jps, jstat, jstatd). The visualgc tool is not included with the J2SE 5.0 distribution, but is instead provided in this unbundled jvmstat 3.0 distribution. The jvmstat tools included with J2SE 5.0, and the visualgc tool included with the jvmstat 3.0 distribution, are capable of monitoring Java applications running on HotSpot 1.4.1 and later JVMs. Only this distribution and a J2SE 5.0 JDK are needed to monitor any version of an instrumented HotSpot JVM.
(..) jstat Experimental: JVM Statistics Monitoring Tool - Attaches to an instrumented HotSpot Java virtual machine and collects and logs performance statistics as specified by the command line options. (formerly jvmstat)
jstatd Experimental: JVM jstat Daemon - Launches an RMI server application that monitors for the creation and termination of instrumented HotSpot Java virtual machines and provides a interface to allow remote monitoring tools to attach to Java virtual machines running on the local system. (formerly perfagent)
(...) visualgc Experimental: Visual Garbage Collection Monitoring Tool - a graphical tool for monitoring the HotSpot Garbage Collector, Compiler, and class loader. It can monitor both local and remote JVMs. (...)

Java Tuning White Paper: Use a benchmark harness

(...) What is a benchmark harness? A benchmark harness is often a script program that you use to launch a given benchmark, capture the output log files, and extract the benchmark score (and sub-scores). Often a benchmark harness will have the facility to run a benchmark a predetermined number of trials and, ideally, calculate statistics based the results of different Java settings (whether the settings changes are at the OS level, Java tuning or coding level differences).
Some benchmarks have a harness included. Even in those cases you want to re-write the harness to test a wider variety of parameters, capture additional data and/or calculate additional statistics. The advantage of writing even a simple benchmark harness is that it can remove the tedium of gathering many samples, it can launch the application in a consistent way and it can simplify the process of calculating statistics.
Whether you use a benchmark harness or not it will be essential to insure that when you attempt Java tuning changes that you are able to answer the question: "Have I gathered enough samples to provide enough statistical significance that I can draw a conclusion from the result?". (...)

Java Tuning White Paper: Use Statistics

(...) Prior to the experiment you should try to eliminate as much application performance variability as possible. However it is rarely possible to eliminate all variability, especially noise from asynchronous operating system services. By repeating the same experiment over the course of several trials and averaging the results you effectively focus on the signal instead of the noise. The rate of improving the signal is proportional to the square root of the number of samples (for more on this see Reducing the Effects of Noise in a Data Acquisition System by Averaging ).
To put Java benchmarking into statistics terms we are going to test the baseline settings before a change and the specimen settings after a change. For example you may run a benchmark baseline test 10 times with no Java command line options and specimen test 10 times with the command line of "-server". This will give you two different sample populations. The questions you really want to answer are, "Did that change in Java settings make a difference? And, if so, how much of a difference?".
The second question, determining the percentage improvement is actually the easier question:

percentageImprovement = 100.0 x (SpecimenAvg - BaselineAvg) / BaselineAvg

The first question, "Did that change in Java settings make a difference?", is the most important question, however because what we really want to know is "Is the difference significant enough that we can safely draw conclusions from it?". In statistics jargon this question could be rephrased as "Do these two sample populations reflect the same underlying population or not?". To answer this question we use the Students t-test. See the Pointers section for more background on the Student's t-test. Using the number of samples, mean value and standard deviation of the baseline population and the specimen population as well as setting the desired risk level (alpha) we can determine the p-value or probability that the change in Java settings was significant. The risk level (alpha) is often set at 0.05 (or 0.01) which means that five times (one time) out of one hundred you would find a statistically significant difference between the means even if there was none. Generally speaking if the p-value is less than 0.05 then we would say that the difference is significant and thus we can draw the conclusion that the change in Java settings did make a difference. For more on interpreting p-values and power analysis please see The Logic of Hypothesis Testing. (...)

Particular importance of Scientific Method in java benchmarks!
Java Tuning White Paper: Making Decisions from Data

(...) It is really tempting to run an application once before and once after a change and draw some conclusion about the impact of that change. Sometimes the application runs for a long time. Sometimes launching the application may be complex and dependent on multiple external services. But can you legitimately make a decision from this test? It's really impossible to know if you can safely draw a conclusion from the data unless you measure the power of the data quantitatively. Applying the scientific method is important when designing any set of experiments. Rigor is especially necessary when measuring Java application performance because the behavior of Java? HotSpot? virtual machine adapts and reacts to the specific machine and specific application it is running. And subtle changes in timing due to other system activity can result in measurable differences in performance and which are unrelated to the comparisons being made. (...)