Oleg Zabluda's blog: April 27, 2012

Oleg Zabluda's blog

Friday, April 27, 2012

Behold, while listening to Аркадий Северный - "Семь Сорок", 7:40, 7-40
Behold, while listening to Аркадий Северный - "Семь Сорок", 7:40, 7-40,

Right foot is scrapped after I climbed up and down the pole all night long, showing off for the girls https://plus.google.com/112065430692128821190/posts/4fUxkiGkuir
https://plus.google.com/112065430692128821190/posts/NZmjRN91nZJ

References (with lyrics)
http://ru.wikipedia.org/wiki/Семь-сорок

Addendum: My father-in-law says it's a so-called "Roman Weight" i.e. my height in cm (174) minus 100 = 74 kg.

42712 | 112065430692128821190

Labels: Oleg Zabluda

In C/C++ data races are undefined behavior.
In C/C++ data races are undefined behavior. Java has a problem: there is no such thing as undefined behavior due to security requirements, so it has to give meaning to them, while not suffering too much of a performance hit ("Java is faster then C/C++", remember?) i.e. allowing allowed reodering, which is the whole point.

Java "solution" is in

Java Language Specification 3d Ed, (SE 7) chapter
17.4.8. Executions and Causality Requirements
http://docs.oracle.com/javase/specs/jls/se7/html/jls-17.html#jls-17.4.8
is a formidable-looking formal specification with fancy set-theoretic math.

Manson, Pugh, Adve: "The Java Memory Model", POPL 05 chapter 9.1.2 "Semantics Allow Reordering" "proves" in Theorem 1 that non-conflicting operations may be reordered by a compiler or runtime.

But in Aspinall, Sevcik, “Java Memory Model Examples: Good, Bad, and Ugly”, VAMP 2007, chapter 5 "Ugly executions", subchapter 1. "Reordering of independent statements"
http://groups.inf.ed.ac.uk/request/jmmexamples.pdf

we read "The ﬁrst example (from [2]) demonstrates that it is not the case that any independent statements can be reordered in the JMM2 without changing the program’s behaviour [...] This falsiﬁes Theorem 1 of [6].

Oops. As of now, this problem in Java spec is unresolved.

This story is retold from Hans Boehm "Programming Language Memory Models: What do Shared Variables Mean?", ECOOP 2011 Summer School talk.
http://www.hpl.hp.com/personal/Hans_Boehm/misc_slides/boehm-ucb10-no-quotes.pdf

Also see
https://plus.google.com/112065430692128821190/posts/KiPn95fTcoa
http://www.hpl.hp.com/personal/Hans_Boehm/misc_slides/boehm-ucb10-no-quotes.pdf

Labels: Oleg Zabluda

Rumors about Google Drive (gDrive) were going on for years before Dropbox was released in Sep 2008.
Rumors about Google Drive (gDrive) were going on for years before Dropbox was released in Sep 2008.

GmailFS from Sep 13, 2004 (http://www.gadgetopia.com/post/2947) was a Linux FUSE userland mountable Linux filesystem, using gMail for storage. It was ported to WIndows in Oct 2005 (http://gadgetopia.com/post/3109). At the time gMail (first released Mar-Jun 2004) was providing massive, unbelievable and jawdropping 1GB of total storage [1], and 10MB individual attachments :-) This forced Yahoo Mail to increase total storage from 4MB to a whooping 100MB.

The name was used for the first time when gDrive 0.6 came out on Feb 12, 2005. It was a set of PHP scripts to store files your gMail account (http://www.puremango.co.uk/2005/02/gdrive_109/).

It was clear for a long time that sooner or later Dropbox will have to compete with gDrive from Google, as well as everybody else in the online storage space. Public statements from Dropbox always stressed not the storage per se, but viral marketing, self-described awesome clients for all platforms, etc... I was always somewhat puzzled why they don't provide other backend storage (Amazon, Microsoft, Apple, ...) as an option, but hey what do I know. Now that the real gDrive is finally out, we'll see if their clients really are that awesome, especially if third-party backends will now be supported.

[1] People actually had hard time believing it at the time, but experiments confirmed it.
http://www.puremango.co.uk/2005/02/gdrive_109/

Labels: Oleg Zabluda

Be afraid. Be very afraid.
Be afraid. Be very afraid.

WR 104 is a Wolf-Rayet class OB mag 13.5 star discovered in 1998, located 8 Kly years from Earth in Sagittarius, more or less toward the center of our galaxy. So what, you say? Some optical measurements indicate that WR 104's rotational axis is aligned within 16° of Earth. When she blows, GRB is going straight for us. And 8Kly is a bit too close for comfort.
http://en.wikipedia.org/wiki/WR_104
http://blogs.discovermagazine.com/badastronomy/2008/03/03/wr-104-a-nearby-gamma-ray-burst/

There are plenty of stars, much closer to us, which are about to go GRB (Betelgeuse being the closest at 643±146 ly) , but this WR 104 is the closest so aligned.

If you are not afraid yet, more food for thought: in 2008, a GRB 7.5Gly away, i.e. before Solar Sytem was born, was visible to the naked eye at mag 5.8 for 30 sec in Bootes at 23:31:09 PDT on March 19, 2008. I could have been looking at it at the time.

http://en.wikipedia.org/wiki/GRB_080319B
http://blogs.discovermagazine.com/badastronomy/2008/03/20/naked-eye-visible-grb/

If that doesn't scare you, I don't know what will.
http://blogs.discovermagazine.com/badastronomy/2008/03/03/wr-104-a-nearby-gamma-ray-burst/

Labels: Oleg Zabluda

Samsung sold 44.5M smartphones in the first quarter – to Apple's 35.1M.
Samsung sold 44.5M smartphones in the first quarter – to Apple's 35.1M.
Samsung sold 93.5M total phones in the first quarter – to Nokia's 82.7M

Samsung plans to double sales of smartphones and tablet computers this year.

http://www.bloomberg.com/news/2012-04-26/samsung-profit-beats-estimates-on-galaxy-phones-demand-for-tvs.html
http://www.guardian.co.uk/technology/2012/apr/27/samsung-apple-smartphone-sales-profit

Labels: Oleg Zabluda

About Me