|
As productivity tools get more plentiful, using the right ones can make or break your company’s retention plan. "It's a poor craftsman that blames his tools"
Is this really where we are these days?
|
|
|
|
|
Enterprises use the Microsoft 365 environment. If I had been the hiring manager in that opening scenario case I would have been breathing a sigh of relief for not having hired "Tony". He would have been a continuing pain in my side. Any one who would rather not make the adjustment to what the employer uses I'd rather not have working for me.
|
|
|
|
|
Time is your most precious asset and slow build is high in the list of developer’s productivity killers. They missed one option, "Whine to your boss until you get a faster computer."
|
|
|
|
|
A former employer was very good at testing, literally taking days to complete. So we started a clean 'release' system rebuild when leaving work Friday afternoon, and it would be complete on Monday morning, ten days later.
So we bought the very fastest machine available, which got all tens of thousands of regression tests complete executed for Monday morning, three days later.
Needless to say: The management never regretted the investment. (Most likely, it paid for itself in a single product release.)
|
|
|
|
|
I'm surprised they didn't get 3 or 5 more of them so they could do nightly test builds by splitting the test suite into parallelizable parts.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, weighing all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
|
|
|
|
|
The main time stealer was a huge (and I really mean huge) collection of regression tests. But for releases, the established procedures required that all of them should be run on the final, fully integrated release candidate, which should be a 100% clean build.
We didn't have any issue with nightly builds: The unit, module and subsystem tests completed overnight. We did have subsets of tests for that purpose. For a release, piecewise regression testing would not be 100% trustworthy.
I fully support the release test regime practiced by that company. The regression tests had grown to a big bottleneck in the release procedures, and it was cured by that investment. There were no similar bottlenecks in other areas.
Additional note: Every now and then a proposal came up to reduce the size of the shipped code by removing test related code inserting conditional compiling directives. Every time that was suggested, half a dozen senior developers each produced a handful stories of where the test code had been extremely useful in diagnosing issues at customer sites. Furthermore, the test regime depended on this code: Running all the tests on a a system built with this code, and then shipping a system without it meant that you were shipping a system that hadn't been tested.
|
|
|
|
|
WTF are these people building? None of our apps (most are largish ASP.Net web sites, circa 2009) take more than 30 seconds to build, and most of them take less than 10, and most of those take less than 5. VS Build performance has never been an issue with me.
".45 ACP - because shooting twice is just silly" - JSOP, 2010
-----
You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010
-----
When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
|
|
|
|
|
I have asked myself the same question, but it happens, so it must be possible.
First time I was really shocked was when I learned that the first, limited release of the software for the ITT System 12 digital phone switched counted one hundred million lines. (I heard rumors that when the project was terminated, the code base had grown twelve times, to 1.2 billon lines.)
A friend of mine worked on the toolset used for this code: The linker ran on a 16 bit machine (this was in the late 1970), but hit one implementation restriction: The table for exported symbols from a single module was indexed with a signed int, limiting the size to 32768 symbols. What the elephant? I can imagine a single module importing more than 32 ki symbols, but not exporting! They made a quick fix changing the index to a uint, but the next major linker release allowed the export of 4 Gi symbols, which should be enough for everyone. At least for a few years.
Also, there was a struct definition - with variants, that explains it - filling 8300 lines. Make a hardcopy printout of that definition, and you have a book of 115 pages!
This was 40+ years ago. The software world have had lots of time to develop even more extreme examples.
When I today hear about extreme build times, the compile and link time is usually a tiny little fraction of the total. Almost all of the time is spent in unit tests, module tests, integration tests. (For System 12, the compile time was in the order of days and weeks.) Pressing F5 is just doing the required compile/link work; testing is a different matter. And it recompiles only what is needed - 'make' is 40+ years old.
Most test frameworks I have seen never discovered 'make'  They rerun the same tests again and again, a unit test is rerun for every module where the unit is used, a module test for every subsystem including the module. All tests are run again, even if not a single code line affecting the unit or module has been changed since the last test run. So for total build times, I think there is a lot to be done on the test part.
Disclaimer:
Maybe there are test frameworks handling such issues nowadays; I haven't really studied them for 2-3 years.
|
|
|
|
|
Seeing as how he's complaining about 43 second builds, I think it's not really a problem. Unless he's hitting F5 at the end of every line.
Quote: Our main solution for NDepend consists of 37 projects, including test projects. Compiling them all takes 43 seconds on my beefy laptop (64GB RAM, Xeon 6 Cores, 2x 1TB SSD). Btw, developing with a powerful machine is essential. Each saving done on hardware will likely costs a hundred more in terms of developer time and focus wasted.
TTFN - Kent
|
|
|
|
|
Elon Musk’s satellite internet service Starlink just got dealt an expensive blow — the company’s currently estimating that 40 of the 49 Starlink satellites it launched on February 3rd will be destroyed because of a geomagnetic storm. "I saw a falling star burn up above the Las Vegas sands"
Your call if you'd prefer the Joni Mitchell version, or Nazareth's.
|
|
|
|
|
A cringe rapper slash Forbes contributor allegedly found with billions in stolen Bitcoin - The Verge
That's a whole lotta nothin', from the heat of a thousand CPUs!
Law enforcement says it was able to seize the 94,636 Bitcoin left in the wallet because Lichtenstein allegedly uploaded a list to a cloud storage and email provider that contained addresses for the wallet that Binfinex’s funds were dumped into, along with the private keys to access them. The file was encrypted, but after obtaining it through a search warrant, the statement of facts says that law enforcement was able to decrypt it and several other files (though it doesn’t offer any explanation of how).
Interesting.
|
|
|
|
|
I was debating posting this one, but I figured it was too ridiculous (even for me). 
Oh, and she’s a “rapper”. The story has everything.
TTFN - Kent
|
|
|
|
|
The whole fact that 3 billion can be directly connected to a single person is pretty ridiculous. And the rest of the story was - chef's kiss
|
|
|
|
|
So much for the idea that blockchain transactions are “untraceable”. That spiral graph hurts my head though.
TTFN - Kent
|
|
|
|
|
Kent Sharkey wrote: That spiral graph hurts my head though.
Not as much as her fashion sense hurts mine!
|
|
|
|
|
Kent Sharkey wrote: blockchain transactions are “untraceable”
Eh? The concept of "public ledger" says exactly the contrary.
You need many wallets and extra technology (like "bitcoin mixers") when you want to make it harder for your transactions being linked to you.
Oh sanctissimi Wilhelmus, Theodorus, et Fredericus!
|
|
|
|
|
Kent Sharkey wrote: That spiral graph hurts my head though.
The person who "designed" that should be locked up in the next cell, charged with graphical abuse. 
Freedom is the freedom to say that two plus two make four. If that is granted, all else follows.
-- 6079 Smith W.
|
|
|
|
|
Blockchain Analytics & Crypto Compliance Solutions | Elliptic
100+ customers across 29 countries/98%+ of global trading volume covered/1 billion+ crypto addresses labeled/66% of crypto volume runs through exchanges using Elliptic
Interesting ... there's something strange in the bitcoinhood who do you call ... elliptic co ... theirs something weird in the blockchain ... blockchain busters...
An affidavit filed against the couple by IRS agents said they spent only a fraction of the stolen funds, some in gold and some in non-fungible tokens or NFTs, a type of A unique digital representation to sell or trade art or collectibles as a unique digital representation. Other payments include Walmart gift cards, as well as payments from Uber, Hotels.com and PlayStation, according to charging documents.
Because the funds were confiscated under a court order, a judge will ultimately decide how to distribute the recovered funds, and the government will seek to return the funds to their rightful owners, U.S. officials said.
“We have cooperated extensively with the DOJ since it began its investigation and will continue to do so,” Bitfinex said in a written statement. The company pledged to “follow due legal process establishing our right to return the stolen bitcoins.”
Bitcoin, the most popular form of cryptocurrency, is computer code generated by publicly available software that allows people to store and send value online. Open source code originated with Bitcoin more than a decade ago and runs on an extensive network of private computers around the world.
The value of cryptocurrencies is usually expressed in U.S. dollars and is determined by public transactions conducted by exchanges; these values can fluctuate significantly. Bitcoin hit a high of around $69,000 in early November before plummeting to around $35,000 last month. It has recovered somewhat and is currently trading around $43,000, according to CoinMarketCap.
Caveat Emptor.
"Progress doesn't come from early risers – progress is made by lazy men looking for easier ways to do things." Lazarus Long
|
|
|
|
|
Just three billion? What is linked to Elon Musk? 
Bastard Programmer from Hell 
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
And it is pretty ridiculous in that case, too.
|
|
|
|
|
Ehr, you anything familiar with how the world works?
Bastard Programmer from Hell
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
Yes I am, and it is still ridiculous. Elon didn't personally make all of the EVs, or pour the pavement, lay the foundations, create the buildings, do the sales, and everything else that was necessary to make what he is credited for. He is simply a figurehead, and the fact that our system gives him a 'value' that is ridiculously above that of almost every other human on earth is downright stupid. What is even more disgusting is that he calls people pedophiles when they turn down his ideas that wouldn't have even helped. Watch Debunking Elon Musk Pt1 (and part 2) for a more detailed analysis of the narcissism at the heart of Elon.
|
|
|
|
|
He personally showed a combat vehicle where the windows got smashed. The figurehead of exploding tesla's? How is that tunnel going that he envisioned?
It not even a competition. My dick is larger.
--edit
Also has nothing to do with BC.
Bastard Programmer from Hell
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
Skyrmions, tiny magnetic anomalies that arise in two-dimensional materials, can be used to generate true random numbers useful in cryptography and probabilistic computing. So stop rolling dice, and start swirling magnets?
Magnets - is there nothing they can't do?
|
|
|
|
|
Is that any more than "Yet another way of generating true random numbers"? Is it anything beyond "different"? Is it faster? Even more random? Cheaper? Easier to include in a chip?
We have had IoT chips with built-in true random generators for years. There are standards for certifying their true random nature. Fr software purposes, these 'skyrmions' are not needed.
I certainly think they should be researched, though. I think we should allow ourselves to study "the true nature of nature" even without having to justify it through some practical application area such as "cryptography and probabilistic computing".
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
