- Writing data to DNA sequences (ST:TNG "The Chase") - Rotating storage (large ceramic rotating cylinders read/write with lasers. In a Alastair Reynolds novel (cannot remember which one). - Same principle as delay lines, but using mirrors on remote planets. It does have to be repeated/refreshed but so does plain RAM. What is the capacity of an infrared laser across 2 light-days?
My boss is all for sharing but have to look into if there is any liability. That'll take a few days. Please email me at isj-firmware AT i1.dk so we can stay in contact.
I'll check with my boss and get back to you (I'm not 100% sure what the license terms were). Out of curiosity: how many GE Zigbee switches/dimmers do you have?
GE ZigBee switches have almost a one second delay between pressing the switch and light coming on.
Yes, the stock firmware has that bug. Although I seem to recall it was the GE dimmers that did that - not the GE switches. It was one of the first thing we fixed in our custom firmware (we deploy them in businesses/hotels).
I agree that a plain rocker-type switch would be preferable for switches instead of the springy-rocker-two-button thing. I think that an actuator would add too much cost. But hey -at least it is not as bad as the touch-sensitive flat glass panales some switches/dimmer have.
What they mean is that there is no or nearly no correlation between first name and last name. So John, Bob, Rob, Randy, Elizabeth, Maggie are all equally likely for surname X.
Of course there will be a weak correlation if the surname is Fleischer then the first name has a slightly higher probability of being Jens, Uwe or Reichard.
The paper is informative. They point out the obvious problems (translation from scripts/orthography missing vowels, but also that many names are actually quite rare. In their dataset 73% of the names only occur once.
They also compare the results with traditional hardcoded rules, and find that neural networks may not be better.So kudos for including non-positive results in the paper.
And if you start out with, say, C++ then you may discover your developers is not developing fast enough.
Both interpreted and compiled languages have their place.
I frequently make prototypes/experiments in python, and then when it comes to making the actual program I may switch to C++ or Java if the performance wouldn't be good enough in Python, or if I need special libraries for dealing with something, or there are some unusual constraints.
CCC-medien has a video about this: https://youtu.be/mGj5Hp354js (in German) I found it quite interesting because the speaker (the one who were raided) tries to also see if from the viewpoint of the police.
The Java serialization feature was fine. It just wasn't meant to be robust against adversaries. Java RMI uses the serialization feature and thus have the same problem. It is fine in trusted environments.
It is a bad idea to use them with untrusted sources.
I didn't RTFA so I don't know if there are other reasons that security for removing the feature.
Java's rmi was not cross-language, so it was only useful if both ends were written in Java.
Corba was cross-language, but had several problems:
- "Design by committee". Designed in OMG, and it wasn't until version 2 that they bothered with specifying the on-wire format.
- No versioning of interfaces. Unless the developers took care not to break things you had to upgrade both ends.
- Cross-platform exceptions didn't work to well. I remember spending a week tracking down an unmarshalling error - turns out the remote call threw an exception and inside that exception was a LongLong and by default the CORBA library didn't instantiate and register a LongLongFactory or something.
- And then there were a fundamental problem (not unique to Corba): pretending that remote calls are the same as local calls. But they aren't. Soon a remote call will time out - something that cannot happen with a local call, and then you have to deal with the complexity of a distributed system anyway. No amount of asphalt or lipstick can hide that.
- And the underlying network (at least with Orbacus) was hidden, so you could not eg. tweak TCP parameters.
Re: versioning: It could have been remedied if they had adopted something like the versioning in Sun/ONC RPC.
Re: local vs. remote calls: It could probably work better if the client-side language supports transactions even on local variables (eg. Progress 4GL).
Something weird is definitely going on with facebook likes and post engagements. Veritasium made some observations: https://youtu.be/oVfHeWTKjag Bottom line: buying facebook promotions can have a negative impact.
Many of the newer DPI and PCEF engines are quite flexible, and can be configured by the customer (ISP/MNO/MVNO). The functionality is neutral and can be used for benign purposes (eg. redirect to top-up pages) or malign purposes (replacing a download). Sandvine is not the only vendor of such equipment - there's also Cisco, Allot, Huawei, ZTE, Procera, Alcatel,...
The article doesn't indicate if Sandvine helped with it or if it was done by the Turkish telco themselves. Given Sandvine's history with the Comcast bittorrent connection reset years ago, I wouldn't be surprised if Sandvine helped, or implemented specific features to facilitate the stuff in Turkey.
Slashdot Mirror
Star Trek is where it's at. See leaked footage of the next movie: https://www.youtube.com/watch?...
You will thank me later. Or not. :-)
plans to make the number function bidirectionally in the future.
Perhaps the company first wants to get some measures in against scammers ("microsoft support") before they allow outgoing calls and SMS ?
blitzgekriegt ?
- Writing data to DNA sequences (ST:TNG "The Chase")
- Rotating storage (large ceramic rotating cylinders read/write with lasers. In a Alastair Reynolds novel (cannot remember which one).
- Same principle as delay lines, but using mirrors on remote planets. It does have to be repeated/refreshed but so does plain RAM. What is the capacity of an infrared laser across 2 light-days?
My boss is all for sharing but have to look into if there is any liability. That'll take a few days.
Please email me at isj-firmware AT i1.dk so we can stay in contact.
I'll check with my boss and get back to you (I'm not 100% sure what the license terms were).
Out of curiosity: how many GE Zigbee switches/dimmers do you have?
GE ZigBee switches have almost a one second delay between pressing the switch and light coming on.
Yes, the stock firmware has that bug. Although I seem to recall it was the GE dimmers that did that - not the GE switches.
It was one of the first thing we fixed in our custom firmware (we deploy them in businesses/hotels).
I agree that a plain rocker-type switch would be preferable for switches instead of the springy-rocker-two-button thing. I think that an actuator would add too much cost. But hey -at least it is not as bad as the touch-sensitive flat glass panales some switches/dimmer have.
Your comment is confusing. Could you please rephrase it in terms of libraries of congress per furlong?
I would like a clarification on your opinion on
The aim of such a development process is agility in software development. The actual process itself often is very rigid (Scrum being one of those),
It is not clear if you think rigid processes are good or bad.
And then that fix goes into the Google silo. The end. Noone else can benefit.
Contribute to OSM and many people/communities/companies can access that fix.
What they mean is that there is no or nearly no correlation between first name and last name.
So John, Bob, Rob, Randy, Elizabeth, Maggie are all equally likely for surname X.
Of course there will be a weak correlation if the surname is Fleischer then the first name has a slightly higher probability of being Jens, Uwe or Reichard.
The paper is informative. They point out the obvious problems (translation from scripts/orthography missing vowels, but also that many names are actually quite rare. In their dataset 73% of the names only occur once.
They also compare the results with traditional hardcoded rules, and find that neural networks may not be better.So kudos for including non-positive results in the paper.
And if you start out with, say, C++ then you may discover your developers is not developing fast enough.
Both interpreted and compiled languages have their place.
I frequently make prototypes/experiments in python, and then when it comes to making the actual program I may switch to C++ or Java if the performance wouldn't be good enough in Python, or if I need special libraries for dealing with something, or there are some unusual constraints.
Angry Flower to the rescue: http://www.angryflower.com/bob...
CCC-medien has a video about this: https://youtu.be/mGj5Hp354js (in German)
I found it quite interesting because the speaker (the one who were raided) tries to also see if from the viewpoint of the police.
You can use 1 exclamation point if a death is involved.
You can use 2 exclamation point if your own death is involved.
When the popup displays, click on the "More Options". Then the big friendly green arrow appears.
You forgot the big green blinking arrow pointing to the green "enable alll purposes" button.
Or that there are 130 third-party companies listed under "measurement". 130
It really depends on what you search for.
When I search for "nvdimm" og "bicycle stand" the first that google shows are shopping results from Google's shopping platform.
Don't expect shopping results for "python substring" or "miniature durian"
The Java serialization feature was fine. It just wasn't meant to be robust against adversaries. Java RMI uses the serialization feature and thus have the same problem. It is fine in trusted environments.
It is a bad idea to use them with untrusted sources.
I didn't RTFA so I don't know if there are other reasons that security for removing the feature.
Java's rmi was not cross-language, so it was only useful if both ends were written in Java.
Corba was cross-language, but had several problems:
- "Design by committee". Designed in OMG, and it wasn't until version 2 that they bothered with specifying the on-wire format.
- No versioning of interfaces. Unless the developers took care not to break things you had to upgrade both ends.
- Cross-platform exceptions didn't work to well. I remember spending a week tracking down an unmarshalling error - turns out the remote call threw an exception and inside that exception was a LongLong and by default the CORBA library didn't instantiate and register a LongLongFactory or something.
- And then there were a fundamental problem (not unique to Corba): pretending that remote calls are the same as local calls. But they aren't. Soon a remote call will time out - something that cannot happen with a local call, and then you have to deal with the complexity of a distributed system anyway. No amount of asphalt or lipstick can hide that.
- And the underlying network (at least with Orbacus) was hidden, so you could not eg. tweak TCP parameters.
Re: versioning: It could have been remedied if they had adopted something like the versioning in Sun/ONC RPC.
Re: local vs. remote calls: It could probably work better if the client-side language supports transactions even on local variables (eg. Progress 4GL).
here (German)
Those also the ones sending out copyright infringement notices for some unknown movie and offering to settle if for ~$100.
So I'm not surprised that things went sideways for Apple - they should be more selective in their choice of law firms.
Something weird is definitely going on with facebook likes and post engagements. Veritasium made some observations: https://youtu.be/oVfHeWTKjag
Bottom line: buying facebook promotions can have a negative impact.
Possibly.
Many of the newer DPI and PCEF engines are quite flexible, and can be configured by the customer (ISP/MNO/MVNO). The functionality is neutral and can be used for benign purposes (eg. redirect to top-up pages) or malign purposes (replacing a download). Sandvine is not the only vendor of such equipment - there's also Cisco, Allot, Huawei, ZTE, Procera, Alcatel, ...
The article doesn't indicate if Sandvine helped with it or if it was done by the Turkish telco themselves. Given Sandvine's history with the Comcast bittorrent connection reset years ago, I wouldn't be surprised if Sandvine helped, or implemented specific features to facilitate the stuff in Turkey.