iheartmemcache

[0] https://www.slideshare.net/SergeyPlatonov/gor-nishanov-c-cor... Interesting to note, slide #11 uses a tokenizer to demonstrate the usefulness of coroutines. IIRC, Rob Pike used a very similar (maybe his was a lexer/parser?) example in '14 re: Go.

[1] http://luncliff.postach.io/post/exploring-msvc-coroutine

[2] https://channel9.msdn.com/events/CPP/CppCon-2016/CppCon-2016...

[3]https://paoloseverini.wordpress.com/2015/03/06/stackless-cor...

--

Side-note : Love it or hate it, the C++ community is certainly moving at a vibrant pace. I don't write production C++ code anymore (and haven't for a long time) but I still go out of my way to watch the CppCon talks. IMO, they consistently produce best quality talks just because you have so many excellent programmers from various industries using select subsets of C++ in all sorts of different ways. The people writing games are focused on making sure they're portable 'enough' to hit all the major platforms, without getting too tied to platform-specific low-level calls, while retaining the necessary performance to get satisfactory models/meshes/lighting/collision detection/raytracing/dozens-of-other-things completed within that tiny 16.33 millisecond gap to complete the framebuffer and swap in the next frame. The academics who are doing vast numerical computations will be talking about their new MPI utilizations (and undoubtedly, next years talk will be re: a boatload of RDMA/NUMA optimizations). People complain that C++ is a mishmash of too many concepts (pun not intended). I.e., you can write it in the "C with Objects" style, or the "I use TMP so much my code is basically Haskell", and anywhere in between - but it's that heterogeneity that ends up yielding such high caliber talks.

I'm guessing Lamport chose not to go the 'fully dependent' route Coq/Agda style (based on his presentation at least) because, well, as he said in the intro lecture - engineers don't really want formal verification at that level but still want strict invariants of their system to be ensured s.t. you can be guaranteed your program will never enter an undefined state nor encounter any undefined behavior as a result entering into an ill-defined state and proceeding to get UB (or 'implementation-specific', ugh) as a result.

What really drove it home is reading the paper Lamport referenced, "How Amazon Web Services Uses Formal Systems" (Newcombe, et al, 2015) along with his RTOS anecdote of decreasing their NASA RTOS' KLOC by 10x while getting stronger guarantees in the process (see: "Formal Verification of a Network-Centric RTOS", Verhulst et al, 2011).

It's also got more than a few similarities with "Fortress". Guy Steele really pushed this programming model while at Sun, but sadly DARPA killed the funding for it. That being said, Amazon has this is at > 14 of their AWS programs so it seems to be accessible and useful enough to the average engineer rather than 3 post-docs locked-up in their academic ivory tower, leaving twice a year only to present at ICFP and POPL.

At the 'formal level', I think the closest analogue to the strength of your guarantee is either Ada/SPARK or QuickCheck. Certainly way stronger than your standard "oh hey, I wrote a unit test in an untyped language, I'm good to go!". (I.e., declare your give me your invariants and I'll make sure your system never reaches a state you don't want).

Worth checking out if only because, hey, Lamport, Paxos. Have a look at Newcombe2015. It's only 8 pages, throw it on the Kindle or iPad, grab a coffee and have a go. What've you got to lose, right?

This is an example of the more modern, detachable design. It's ready to use in under a minute.

As a child growing up in the suburbs of Boston, we had something similar in our backyard. It would come out with the summer patio furniture and return the shed when autumn hit. Not only it more eco-friendly, but the long term damage of high-heat is pretty well studied in the scientific community[1].

I was actually interested enough to test this out on my own. During the regular Brooks Brothers 2 for $100 bi-yearly sales, I picked up two identical blue oxford cloth button-down shirts (of domestic manufacture, circa 2012). I weighed them to make sure the oz/in quality was comparable, along with a pretty thorough defect analysis to inspect the quality of the seams, how well the buttons were anchored (they both had anchor points), qualify of cuff/plackets, etc. They were more or less comparable in quality, but I selected the one with more flaws to be used as the 'air dry' (I took a Sharpie and scribed with a red x to denote the 'air dry' to maintain consistency.)

Both of these shirts were washed in the same unit, with the same detergents and had uniform wear patterns (amortized over time at least). They'd get ~3 wears in the winter, and generally 2 in the summer, followed by a wash in cold/cold with on 'medium' agitation. I wish I had pictures to document the evolution, but after ~6 months the color fade became was pronounced to such an amount that I could identify between the two. This fade continued in more or less monotonic ln() fashion as a function of time.

The structural failure occurred on the 'dryer' unit around 3.5 years into the purchase. It failed at the at the elbow, though curiously not due to fabric wear fatigue in a stress-loaded position (i.e. the conventional 'blazer hole'), but the seam that bound the sleeve together[2].

Obviously there's no rigor to this anecdotal story. But any article of clothing that I've got some sentimental attachment to will get air-dried just to prolong the longevity. For those who are curious, the 'air dry' experimental unit was retired after ~6 years of use due to structural deformation[3] of the experimenter.

====

[1] https://www.sciencedaily.com/releases/1999/08/990831080157.h... [2] I.e. construct a paper cylinder, slide it over your lower arm, and cut it in from your elbow towards your hand. [3] I.e., I put on ~25 lb of the indolent variety between 24 and ~29. I went up a jacket size (36R->38R) and 2inches of waist.

Likewise, fields advance quickly. I can grok how an z80 or 6502 works from NAND to Tetris, but even a mediocre second year grad student would wipe the floor with me. I, too, went pretty far down the road of mathematics, but watching MSRI lectures from the last few years leaves me struggling to keep up, in the field (algebraic topology) where I once felt comfortable. If you don't keep up with your field you're going to be lost.

The reason I think the 'black magic' trope keeps on being bandied about is because most people reading the articles describing ImageNet et al just don't have the background necessary to grok it[1]. If you had asked them a year ago what the convolution operator was, they'd have scratched their head. When they try to go and read that ImageNet paper they'll be left even more confused because the last time they thought about linear algebra was in their freshman year of uni. It'd be analogous to trying to write some computational fluid dynamics modeling software after not having taken/not touching diff eqs for a decade.

[1] This isn't to disparage those who didn't- everyone has their domain of expertise. I'm just trying to emphasize why the conception of 'black magic' exists. It's quite simple - when one has a tenuous grasp on the foundational knowledge upon some theory is built, you will have difficulty learning abstractions built upon said foundations.

The benefit of medical-grade "AI" (in particular, multi-layered convolution neural networks) is that, given an aggregate of information (say, if you equipped every derm and oncologist with high resolution cameras and a set of parameters to standardize each datum), a trained professional[1] would be able to use that corpus as a very useful resource (used, obviously, in conjunction with their formal training and years of medical experience).

That being said - this is a question you should really be asking those who practice medicine or are actively in research for a living. Go pick up the last years issues of Nature Methods to see what problems they're encountering, and which technological gaps[2] (if any) they may have where YC AI might be applicable.

----

[1] Needless to say, this is something I'd be very reluctant to release into the populace's hands, lest you have someone whip together some node.js backend and React iOS app :: "Do I have Skin Cancer? $10 to find out!". One shudders to think what sort of hysteria might follow.

[2] You would be surprised at how technologically adept some of the members of those research teams are -- within a month of DeepMind making the headlines on our tech blogosphere, I was seeing RNN being applied to their industry.

I've used Squeak Smalltalk and DrRacket in the past. As someone who finds slime/swank/sbcl in Emacs a pleasure, I can only imagine how great some of those other environments can be.

That being said - let's take a field which changes fairly rapidly, all things considered - medicine. You have two forms of changes which can occur within your data model -- (a) definition changes where your schema stays the same effectively and (b) paradigm changes. Let's say you're with a patient and need to code in a ICD[1] value for whatever malady he has. A crude analogy would be a comparison between the Dewey decimal system and the ICD. As the ICD evolves, new codes might be added but the fundamental structure remains the same. So, a new database paradigm TotallyAwesomeRuleBasedImmutableDB pops up, and you just use the pre-existing code for "Databases" (say 20) and appropriate the next available integer in that set for TARBIdb(lets say, 73). No problems there. Essentially, we just inserting a new value (73, or added a new "smaller thing") into the set of Databases to universe U_1 by without violating any invariant properties. Consider (a) a non-issue.

The problem occurs when you need to address the issue of your patient being an alcoholic. Previously, this was referred to as trunksucht (loosely translated as: "dipsomania"). This would fall under the 305, 303, 291, or 980 codes depending on if you were say, a non-drinker with no psychological cravings for alcohol who pledged for a frat, drank too much and developed alcohol poisoning as a result of that. If you were a daily drinker exhibiting physical withdrawal symptoms, you would be coded from an entirely separate section. From the ICD-9 to ICD-10, the WHO reclassified this entirely (such that either patient would be coded in to an entirely new category 'F10').

This breaks a lot of things, from the front-end GUI which would do validation with a regexp [0-9]{,3}\.[0-9]* or some such which is trivial enough I suppose, to the stored procedures which then have to get vendor approved schema changes, all of which will be rolled up and sent to your IT guys who have to go through tedious QA to make sure everything works. (And I mean literally every modal dialog, every validation, every external XML-RPC call to bill out your patients insurance company, Universe U_2).

I've read Codd's paper, the entire ANSI SQL11 standard, and have a fair grasp of both major vendors. There is nothing in there about gracefully dealing with these situations. I suppose one could argue that an RDBMS' purpose is solely to act as a data-store while retaining availability and data integrity -- that is, U_1 is all your Oracle instance cares about (manage your tablespace's set of schema objects and the interaction between U_1 <--> U_2 is completely out of the scope of it's set of responsibilities). Maybe so, but when Oracle added the XMLType with full XPath selector capabilities in 9i (give or take?) they certainly co-opted the responsibility to deal with changes of a structural (that is, "of type 'b'") nature. Dynamically typing, weakening of type systems, writing to an 'duck typing'), serializing via the Protobufs methodology, etc are all just band-aids; we need a surgeon to go in there with some surgical sutures.

[1] https://en.wikipedia.org/wiki/ICD-10

From an economic POV, I should be outraged that I'm continuously paying money for services that states that consistently run significant net-negatives (e.g. Alabama). I'm fine with that[1] - it's a component of the nation-state, and it provides allows states like Alabama can to keep their highways operational and their schools funded. I accept this as being a member of a modern nation-state so that my fellow countrymen with less opportunities can have a shot at a better life.

There are some institutions like the Corporation for Public Broadcasting that take tax dollars so I can watch the PBS Newshour, even though it doesn't generate revenue. I am a-okay with the National Endowment of Arts because it produces 'culture', which we'd both agree creates value, despite us not being able to put that numerical value into a spreadsheet.

I'm not a German national but I spent a few months in Berlin. It started out as a stint for CCC (back when it was in Berlin not Munich) and I extended my stay because I met some incredible locals. I'm quite fortunate to have traveled extensively enough for work and pleasure to be jaded. You travel enough and you end up meeting the same types of people doing the same things with minor (but very extensive and often very subtle, primarily cultural) differences. Some guy you did a 3 week stint for in in Joburg will have the exact same mannerisms as someone Nicaraguan you end up having a chat with at the airport bar waiting for your flight out of JFK. I say all this as a preface to establish this: I'm not really very much impressed by the novelty of this city versus that city.

However, Berlin is one of the few cities that provided a certain unique je ne sais quoi ('feel'? 'ambiance'?) to it, that I'd imagine must have been similar to being in Vienna during the 1920s for philosophy, Greenwich Village NYC for folk-music in the 1960s, or Silicon Valley during the hay-day of NatSemi. Not just in tech, mind you, but take the local music scene there, absolutely spectacular, as is the broader category of arts. The city just overtakes you if you let it, and makes you want to create.

Why do you think the population is increasing? I couldn't say for certain but I'd argue it's because others feel similarly. I'm not sure what ingredients or at which proportions to make that specific cocktail, but I'd argue when you have it, it's worth a few extra greenbacks, pounds or Euro to sustain it.

==

[1] Other than the large percentage which goes to fund the everlasting 'war on terror' and all of the beneficiaries of the military-industrial complex. Let me ask you, if there's a 'war on terror' or a 'war on drugs', Who do I sue for peace if I want to surrender and start negotiating resolution terms? How do you win a war against a concept?</rant>

#2 is the whole risk aversion thing. Any platform migration is going to have political risk attached to it. Nike won't move away from Oracle even if Postgres can do everything it can line for line because paying the $200k maintenance fee each quarter is a cheap recurring cost compared to the political risk of a failed project.

That's not even factoring in #3, the little features here and there which are value-adds to the platform and not part of the standard. I've inherited projects where my firm offers to support O&M for some poorly written ASP 1.1 Classic software. We'll go in, patch some business-logic bug, and get a phone-call later that afternoon asking to rollback our fix. Turns out their users had integrated bugs into their workflow such that when the software operates "properly" (i.e. 'to spec') it actually breaks things (i.e. stops operating in an expected fashion).

I'm not in medicine (oh thank Christ) but I come from a family where I couldn't ignore the the industry if I wanted to. (Literally every male for 3 generations on both sides have worked directly in either medicine in a university capacity after their PhD, actively practice(d) medicine, or more commonly done both.) My mother's an RN, and my sister's a BS/LPN who's one of those "do gooders" who jumps around (on her third graduate degree now) working usually neuro AC or OR full-time concurrently. So I've heard it from the board-level to the overworked resident side to the 'thank god that fellowship is over' and the 'just got off a double and had three GSRs and a handfull of standard MeOHs', from the the early EHR adopters (interestingly enough, B&W/Partners in Boston) side, the pen-and-paper side, to the 'in between'. The 'in between' is the worst for every one, at every level, bar none, unless you're the one on the steaks-and-strippers sales side.

The limbo period is where absolutely everyone is frazzled because things don't work the way they used to. That's fine for us since we design the damn things, if the scheduling API changes we can just use ...anything really... to make that new JSON invocation or XML-RPC call and Bob's your uncle. But everyone can't be expected to "just pop the hood change the carburetor's flow by 3cfm"; they just want their damned car's engine to turn over so they can use the vehicle. So everyone else is fighting new apps and new workflows while still having to make sure that you're still maintaining an SLA of literally 24/7 operational functionality in certain departments with maybe 3 hours a quarter of scheduled yellow-time for upgrades on 3AMs of the first Sunday each month.

Let's start with getting paid, since everyone needs a paycheck, right? The new system dictates that all employees have to badge-in with new RFID's at certain points (entrances, exits, secure regions, etc). In theory, great idea. Part-time employees salary now gets automated plus now we can ensure better operational security (why is Dr Jane, 9AM-6PM, internal medicine, badging into the OR, 4 floors from her office 5 hours after her shift, where all the Demerol happens to be kept...?). Any non-salaried employees pay-check gets tied into the badge-in/out and literally 70% of the staffers forget to do this, since they never had to before and after lunch they're just walking back in as they normally would. Jon the Janitor pulls up his Bank of America statement and is wondering why last weeks check (bi-monthly, with one pay-cycle close period for adjustment/payout -- so some of these hours, remember, are from potentially 5 weeks ago) was for $90 instead of $2400. So he calls HR to get the money he worked for, and Oprah who's been a great HR employee tries to figure out where the 'alter' feature is and after 5 minutes of "bear with me, sorry, this is a new system...(awkward pause)..." they finally figure out how to make adjustments to the hours (turns out the a manager 2 levels up has to authorize the labor change, since it's past the adjustment business rule of +/- 10% of the hours). AP is slammed with a 40% increase of workload since all those checks have been sent out and we're already in to the second week of the new cost-accounting period. So everyone is frantically calling in 5 weeks after the quarter closed as to why last weeks check was for $90 instead of $3400. Having fun yet?

So Tim works the support-desk phones making sure Derm's scheduling is taken care of. He's spent 20 years making sure all those follow-ups are properly scheduled, all the new patients coming in have the proper primary referrals, etc. Not a glamorous job but he's good at it. Jane's in her 50s and been seeing Dr Eczema for 8 years. She's calling to schedule her yearly just to make sure that mole on her neck which was 2 of 5 on the ABCDE ("looks alright now, let's keep an eye on it though") hasn't developed into a malignant growth. He tries to pull up her file (there's about a 75 second delay there as he fumbles around with the new system's UI and "no no, that's not your subscriber ID, that's your patient ID, your member ID is 8 digits and begins with 40", he finally gets to the right modal menu, and has no idea to even see if her doctor has any availability on the Tuesdays mornings she's got some free-time to take off from work. (Before this, he's would have always just told Jane to "please hold.." and punched in the extension number for Dr Eczema's PA, Mr Clockwork to see if there's availability.) So now he's literally iterates through every possible drop-down, for every 15-minute interval Jane can take off work (9AM to 11:30AM, 10 slots), for the next 6 weeks (takes say 4min30sec per patient). It ends up taking him five minutes, but he gets really good over the course of 8 months at clicking buttons (clocks that down to just under 2 minutes, boom!) before a co-worker tells him there's an "Availability" modal hidden 3 sub-screens in that lets you access Dr Eczema's full calendar. (He was sick on that training session day.) Tim's ecstatic.

Don't get me started on the the custom iPads that the CIO thought "we gotta have, it's the new thing!" which no one uses since the GUI is obtuse and every room has a desktop which is far easier to take notes while talking to your patients. 6 months later, 2k iPads are gathering dust and everyone just uses an RDP session into the Win2k12 server since it's way way easier to take notes/make referrals/write scripts. Don't get me started on actual selection process, as board-level CIOs' wife happens to hold 200k shares of pre-IPO stock she scored 12 years ago that's worth a small fortune now. He conveniently omitted that fact, and rather than recusing from the selection committee was actively pushing for it. (True story.)

Now the CIO is evaluating 'hmm should we go to Cerner?' because we have 2k iPads sitting around doing nothing, some 60 year old absolutely-brilliant neurologist who's responsible for a ton of patient draw (as such, a lot of political clout) can't figure out how to use the damned thing. (An apt comparison would be "ever tried to help your grandparents use email during the hotmail-era?" for those of you who remember that sort of thing.) [Basically the equivalent of moving from SAP to Sage500/Dynamics/other-ERP which also happens; the Mayo Clinic transitioned from EPIC to another EHR while my uncle was there, all sorts of fun I heard!]

There are all sorts of issues on the logistics "who gets what money for what and when". E.g. you have changing insurance issues (i.e. OK, so this patient with this internal medical record saw Dr Foo in dept Bar, which we can internal cost accounted at $x, let's get this invoice for $y out to... (oh damnit, ok who is their health insurance company, let's ping the state's open exchange since this patient gave us their MassHealth ID number[or whatever appropriate exchange your state has if they opted to take the ACA federal funding] and see who their insurance provider is so we can bill this out). Single-payer would be a blessing in reducing complexity!

If you've ever wondered why some cities drop 12 million dollars on a new HR platform only to have it fail 18 months in, it's because information management is hard, regulatory reporting burdens are high, and transition costs (both the political stake the primary project manager who initiates the change-over, as well as the actual quantifiable monetary costs) are astronomical. ATC still uses software designed in the 70s where their web interface is literally screen-scraping IBM CISC software. It's horribly inefficent, but it's been battle-field tested.

Imagine if cardiologists all of a sudden went 'value-based free market' in their pricing. With all the retiring boomers in their waning years, I could very easily see 6 hours of labor for a staff of ~6-8 in the OR performing a bypass being billed at $100k.

If you want a completely 'free' (as in GPL) cell phone experience, you can setup a OpenBTS transmitter and transmit at the 900mhz range which is commons property. To stay legal in the US, your antenna has to put out less than a watt, but the setup allows you to even use off-the-shelf phones and trunk into normal phone lines via standard POTS software. Your device would have to be something a-la http://alumni.media.mit.edu/~mellis/cellphone/ (just a janky setup, but just a proof-of-concept -- you can patch together components from DigiKey pretty easily these days; if you want free-silicon, I think the closest you're going to get is https://en.wikipedia.org/wiki/OsmocomBB or maybe some soft cores, but if you're actually going to take that soft core to tape-out, you're probably going to be running 6 figures just for masks...)

I have extreme reverence for Bill Nye as a voice of rationality[0], so when I heard him on a talk show nay-saying nuclear power with his prime argument of 3 failures against ~400 I poked to see what sorts of impact it had. At first glance, a failure rate of of almost 1% is absolutely insanity. Things have significantly changed, however, since the design of Chernobyl. Control technologies had progressed in modern (3rd and 4th generation) nuclear power plants, which are safer and cleaner by quite a bit compared to what the 60s hippies were riled up against. (In fact, there are 3rd generation plants that can be powered solely off of the waste-materials of ~7 older plants, providing an auxiliary benefit of consuming spent-fuel rather than having to can->seal->bury). The important thing to consider is that as technology progresses, we (humanity as a collective) should revise our expectations. After all, we don't factor the failure rates of Wright brother era endeavors into Boeing numbers.

Either way, seeing that segment got my curiosity fired up enough to begin some inquiry. (Standard caveat of "everything that follows is armchair"; this isn't even close to the field of study I'm well-versed in. The closest I get to energy professionally is low-voltage/low-amperage li-ion power control, spun such that units pass UL/CSA.) Were his fears founded/relevant to modern nuclear power sources? Were his fears a product of him living through the Cold War as well as living through all 3 of the major disasters that made his fears more emotional than my own? (I.e., he's in his early 60s and lived through the media coverage of 3 mile, Chernobyl and Fukushima[2] which would have impacted his opinion; I'm about to hit 30 and only Fukushima impacted my perception so my fears might be assuaged easier than his.)

I turned to full-cost accounting (from drawing board to decommission; factoring materials sourcing and the ecological impact (e.g. Aluminum takes a lot of power to produce); factoring in transport of sourced materials; maintenance (operational/preventive and modeled failure red-tag fixes); waste disposal) makes nuclear look somewhat appealing, even compared to PV/wind (typically deemed it's closest competitor). As one would expect PV/wind/hydro/thermal/nuclear all score about the same (give or take 50% for all of those technologies), while conventional non-renewable hydrocarbons operate within the next order of magnitude up.

So where does this leave us? Well, certainly there's agreement that fossil fuels/hydrocarbons should be phased out sooner rather than later[4,5,6]. From the damage done to the coal miners lungs, to the carbon impact they're about an order of magnitude less efficient (half-an-order with newer processes to reclaim heat excess and spool up turbines, not unlike using your cars' exhaust volumetric flow to spool up turbos as a form of auxiliary power).

Similarly, those super-tankers that get our iPhones from Foxconn to California off of diesel are doing quite a bit of damage[7] as well. Russia is still using nuclear power for their ice-breakers in the Arctic without incident. I'm not sure how close to catastrophe those units are, but all of the active US Navy combatant submarines are nuclear powered. I'm sure those sailors are thankful they don't have to drink water tainted with diesel as they did in years-yore. I wonder if anyone has conducted viability studies re: utilizing the same power mechanisms that fuel the sub props for those super-tankers (genuinely not sure - if anyone has more information, hit Reply).

Apples to apples (e.g. they both happened within a few years of each other) - the Deepwater Horizon had a catastrophic impact on our ecology. I'm not sure how bad the long-term ecological impact was compared to Fukushima, but I know after a "BP cares" PR campaign, we got right back to business as usual -- we're still pumpin' that black-gold - no one shutdown any hydrocarbon based operations as a result of the catastrophe. The only reason I can think of Fukushima having such a large impact is the historical weight "nuclear power" is burdened with. Someone needs to bring in a PR firm and come up with nuclear's equivalent of "clean coal".

==

[0] Shills on Fox News show up as guests frequently and try to attribute climate change to 'natural cycles in the world' using cherrypicked data (say: CO2 statistics from a selected 7 year period). Nye elegantly anticipates these arguments (not too hard, I suppose, since there are a finite number of them) and then contextualizes them (e.g., bringing out a graph showing a 100 year period). Go watch it on youtube, it's quite entertaining and a very elegant way of argumentation that's accessible to the "Glen Beck" demographic [using pretty charts instead of a chalk board].

[1] https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit...

[2] http://spectrum.ieee.org/tech-talk/energy/nuclear/fukushimas...

[3] https://www.bis.doc.gov/index.php/forms-documents/technology...

[4] https://blogs.scientificamerican.com/the-curious-wavefunctio...

[5] https://en.wikipedia.org/wiki/Cost_of_electricity_by_source#... - Note this is consumer cost, not "full-cost".

[6] https://www.brookings.edu/blog/planetpolicy/2014/05/20/why-t... (Standard full disclosure any time I cite a source that's not purely academic/where conflicting interests may exist: Brookings is a "liberal" think-tank. Even though I agree with most of their publications, I don't think they should have an .edu any more than the Cato Institute should.)

[7] https://www.quora.com/Is-it-true-that-the-15-biggest-ships-i... This question was posed in response to a Daily Mail (UK) article that was sensationalist-journalism. The Quora response is of far better quality and addresses the nuances with significantly more insight than the original article.

[1] https://www.cse.unsw.edu.au/~cs4161/lect.html

I.e., LAPACK/BLAS benchmarks are just really big linear algebra matrix problems, so obviously your pre-fetch and branch prediction performance will be significantly better since you aren't dealing with interrupts, locatedb, or Windows DCOM events firing off in the background. You have a huge set of matrices with a very predictable set of branches, fetches, and decodes, so obviously your CPU can optimize for that load, you're just paying for it in latency on the back-end (RAM fetches are the new disk swap ;)).

On all those benchmarks, (i.e. your standard LU matrix decomposition which previously was the basis of the LAPACK benchmarks, though things might have changed in the ~10 years since I've really looked at things) isn't CPU-bound anymore, so of course your instruction-per-cycle load on the CPU isn't where you'll be bottlenecking (and hasn't been since "let's avoid floating-point operations and just use static look-ups instead since we don't want the 10x cost of using the FDIVP instruction!"). Your processor can very easily anticipate from where in that sparse-matrix your next data fetch is going to be. It's the cost of that RAM fetch[1] going along that copper trace which is going to be where you're going to bottleneck on any heavy numerical computation.

The power consumption on your CPU might drop a nominal amount which is great for those marketing white papers, but for a numerically heavy load, you're paying just as much (in total power consumption per 4U in the data center, total heat generation/dissipation within the case, and total processing time) on the back-end for those fetches.

[1] https://i.stack.imgur.com/a7jWu.png (I normally cite academic references, but this is 'good enough' to convey my point, I hope).

Rhode Island was founded because the Puritans quite like what Roger Williams had to say. Ben Franklin was noted of making a remark about there being too many Germans. The Potato Famine brought Irish emigrants and with it slurs ('mick') and hatred against that demographic. The Southern Italians weren't considered white when they came around (also had their slurs coined, 'dago, 'wop', 'guinea').

A sudden (or continuous, but seemingly overwhelming) influx of an overwhelming amount of people who are "new" upsets the ecosystem. If me and, say 30 of my (relatively) leftist, mostly-holding-graduate-degrees friends (and their spouses, etc) all moved to a small-town in West Texas, surely we'd be looked upon as differently. It wouldn't be because I had brown skin (which I do, and there might be a little bit of 'hmm that fella looks kinda strange' at first). I can tell you from contracting all around the world from rural Alabama to Beijing, that the reason my buddies and I would be singled out would be our mannerisms, our speech patterns, our dress, our modes of communication, and what we consider polite. (I.e., when I'm in NYC, I may see a friend sitting having a coffee al fresco. He's busy on his iPad, I'm late for a meeting - I'm not going to stop and say hello. Contrast that to someone who might take that as a snub and potentially harvest resentment for years.)

This allows me to segue onto a tangent and make an important sociological distinction (it's somewhat relevant, I swear).

Lets say grew up in a small town in Alabama where your pa and all his kin have been for damn near 3 generations. It's completely reasonable to be slightly off-kilter when an aggregate of people disrupt the stability of your living situation. Your memories were formed in that town, you're invested in all sorts of ways in that town - emotionally, financially, you name it. Your memories were formed in that town. The money you made from mowing lawns when you were 12 (do kids even do that now, or do they just Instagram and yell at their parents?) was spent at that local ice cream shop. You had your first kiss in the back your dad's car after you and that girl shotgunned a joint in the school parking lot cutting class. There's still a blood stain on concrete of your parents houses driveway from when you hit on that jock's girl and he came around and broke your jaw (OK, that one didn't happen to me, but you see what I'm getting at, I hope).

That doesn't make you inherently racist, it makes you upset at the prospect of losing some stability you used grew up with and expected (however reasonable or unreasonable those expectations were) to have. Now, if I come around with my buddies, we buy your parents house, knock it down and throw up some ugly McMansion, get a 30 year lease on that ice cream shop's property, knock that down and replace it with a raw-food vegan or fair-trade we-only-serve-black-coffee-here cafe. You'd be upset at the culture shift and not because some brown dude did it. There are certainly racists out there, there's been this media depiction of any Trump supporter being one. This is problematic because it not only does it mask the real issue (there are active neo-nazis out there, burning crosses, and sitting around hating everyone as their Friday recreational event) but it divides the country even further down political party lines.

A good litmus test between "professional" and "premium consumer device" is 1: do you have a direct line to a domestic tech who can address your issues (i.e., generally not having to waste your time with a level 1 service tech, but directly reaching someone who is at least semi-competent and capable of authorizing work orders and/or warranty claims in the ERP) or an option to buy an SLA, 2: does the company offer a mil-spec and/or pre-authorized GSA listed items, 3: does their financing dept offer 3 (or 5, depending on how your organization decides to do their depreciation) year equipment lease schedule ?

[1] https://www.lenovo.com/images/products/services-warranty/dat...

Their site has amazing documentation. I personally found Qt even in the KDE2 days miles better than React/Redux and non-Smalltalk MVC's. Since then it's only gotten better. Check out the examples here[1], the diving in section here[2], the book here[3], Stackoverflow has an overview of the whole system here[4].

You emit signals. Slots consume them. QObject::connect links the two. Queues are just queues in the traditional MQ sense. You don't use them in for, say, local GUI controls since it's generally a 1-1 unidirectional link (and if it's bi-directional, although I rarely need to do this, just ::connect( a,b,SLOT(),SIGNAL() ) then ::connect( b,a,SLOT(),SIGNAL() ). If you're still having problems read this[5]. If you have issues, that blog also has a preceding page which has a comprehensive analysis of the underpinnings of the entire slot/signal system in QT. If you still have problems, get on IRC and someone will give you a hand. If you're having a major issue, contact me. If it's a quicky I'll lend a hand, if not I know a few absolutely spectacular HN engineers who I'd recommend in a heartbeat.

I'd be real curious as to what's you consider 'messy' about signals/slots is you're having re: slots/signals (I guess since I've used Moose's MOP with Perl 5 and CL they sort of came naturally to me). I have no affiliation with Trolltech/QT/Nokia/whoever-owns-whatever-now at all; I just think C++ is real easy to mess up, libraries are even easier and they did a way-above-average job with their code-base (not to mention, documentation, ecosystem, etc).

The setup you want to use for cross-platform is Qt5.x + QML 2.x + QtQuick 2.x (aka QtQuickControls 2 in newer versions). There's licensing issues on some of the fancy-schmancy widgets but most of the things follow LGPL, so just dynamically link, follow the rules, and you won't get sued. If you're working on a legacy codebase (gcc2.95.3 hollaaaaa), don't try to force new-style QT into an old project, or your experience will be similar to @bwoj. 5.6(or there abouts) and up is where QML started to shine.

N.b. you don't have to use QTcreator. I find it satisfactory but that just might be years of Stockholm Syndrome. Most people coming from the C++/MFC (or C++/WTL/ATL/COM/WinAPI world) are more comfortable with Visual Studio. No problem, the 'standard' procedure is just to do what you would do with any .sln and split GUI into another project.

----

[0] https://www.qt.io/qt-essentials-qt-quick-for-c-developers/

[1] http://doc.qt.io/qt-5/qtquick-codesamples.html

[2] http://doc.qt.io/qt-5/qmlapplications.html

[3] https://qmlbook.github.io/assets/qt5_cadaques.pdf

[4] http://stackoverflow.com/a/19837895

[5] https://woboq.com/blog/how-qt-signals-slots-work-part2-qt5.h...

Here's a guy who runs a mini-vineyard in CA who saved almost half a million gallons of water in one season for his grapes using pretty trivial means[1]. I bet if he incorporated some recent 'best/alternative' practices he could get that significantly lower too.

The good news is universities (particularly Australian unis[2]) have conducted pretty thorough analysis w/r/t efficient watering for healthy plant root systems. You have lay new sub-surface irrigation drip system with textile enclosures[3] which is an initial expenditure, but you end up saving a boatload of money on water in the long run. (Drip irrigation[5] in general is alright, but you have tons of water loss via gravity[4] as shown here.) "Permaculture"[6] is a DIY-sociological movement of farming by 'efficient/sustainable/whatever' means.

The bad news is the last time the US had a radical shift in farming was when farmers beat their land so hard we ended up having the Dust Bowl. Conditions had to get so bad that you'd get 'dust days' like 'snow days' from school/work for people to change their farming habits. It wasn't until FDR appointed Henry Wallace(unsung national hero) as Secretary of Agriculture that things started to change. If I could buy futures in water 50 years from now, I'd put my entire savings in it. The botanists and horticulture scientists have made massive advances since the 1930s. We can sustainably farm on land every season if you don't plant corn every dang season and then pump nitrogen in there to make up for the lack of nutrients.

Grey water[7] can very easily be treated and reintroduced into your local garden. We can even use refuse from septic tanks, treat it on-site and re-introduce into your your lawn/garden to produce 100% safe edibles (subject to your local regulations; also, if you're on any sort of medication such as antibiotics, you don't want to introduce that into your ecosystem). No land? No problem. You still can save water. Get a dual-flush toilet[8] which uses half the amount of water to dispose of urine.

I'm not a 'save the earth' kind of guy. The planet was here long before humans and will still be around regardless of all disasters we can possibly throw at it; that's a certainty. Will your grand-children have congenital defects from rampant pollutants, poor air quality, and tainted water is the real question. (I'll end this on a good note - the good news is reforestation can be done in a generation or two![9])

--

[0] http://www.lindsay.com/stuff/contentmgr/files/0/6ab577c195d6...

[1] http://hackaday.com/2015/12/11/vinduino-full-irrigation-with...

[2] http://www.unisa.edu.au/Research/CWMR/Students-and-researche... - These guys are doing some interesting work. Most of the Aussie unis are actually. U of Kansas is doing some great work too.

[3] https://en.wikipedia.org/wiki/Subsurface_textile_irrigation - They mention PE and PP but there are alternative materials available which aren't hydrocarbon based

[4] http://www.intechopen.com/source/html/45153/media/image4.png... Effects vary on type of soil (and subsequent density that determines the rate at which gravity pulls water away from the root system), the micro-culture that exists, how water-hungry your plant is, and a boatload of other factors...

[5] http://www.sswm.info/category/implementation-tools/water-use...

[6] https://deepgreenpermaculture.com/diy-instructions/how-to-bu... Water rights will vary from state to state, especially in the Mountain/Pacific zones. It gets as extreme as "you're not allowed to collect rain-water on the land you own, sorry" which is insane because such legislation doesn't even apply to oil/gas/minerals. If you hold full title to an estate (i.e., you didn't just purchase surface rights, or a split claim) and you find diamonds on it, you are issued full rights to it.

[7] "Greywater Reuse" - Amit Gross, ‎Adi Maimon, ‎Yuval Alfiya, 2015

[8] https://www.amazon.com/K-3987-0-Wellworth-Two-Piece-Round-Fr...

[9] https://www.ted.com/talks/shubhendu_sharma_how_to_grow_a_for... I hate TED talks, but this former Toyota salaryman learned about biodiversity from some 80 year old Japanese guru, applied the Toyota manufacturing techniques to make the system efficient, and is doing some amazing work with basically no funding behind him.

Auditors were supposed to make sure books were GAAP-compliant. That was their entire job. Even they can't/won't keep up (RIP Arthur Anderson). And that whole "you have to file an 8-K with teh SECevery 3 months with the SEC" is limited to publicly traded financial institutions within the United States[2].

To make an analogy, the law can barely keep up with tech (see: amorphous definitions of 'data collection', 'meta-data', is your cryptographic keyphrase protected as a fifth amendment right, is PGP a "munition"?[3]). Financial markets are innovating at just as quickly a rate as we are.

So, in a modern day, how do we prevent 'cooking the books'? I assert that the problem is lack of transparency. Here's an idea that probably has many major issues as I haven't ruminated on it at all, but let me give it a shot: require all transaction data[4] to be independently audited by any stakeholder at any time. As a stakeholder who is keeping my retirement money with this institution I'm highly motivated to ensure that whoever is managing my money is doing so in what I deem to be a proper fashion. Access to their books (obviously with the account information/PII scrubbed), maybe after a business week of buffer time (to allow for some padding for any tactical content that may be in the current positions) should be perhaps be made available.

My second assertion is this will not only keeps the institution obligated to be 'honest' with their reporting as obligated to their stakeholders, but has an auxiliary benefit of incentivizing their competitors to act as very motivated auditors as they scrutinize their competitors books to uncover any malfeasance.

Since regulatory bodies can't keep up, _ensure that all actors within the market actively keep each other honest_. Sure, they can collude but collusion is far easier to detect and prosecute than internal book manipulation.

----

[1] Subsidizes to under-produce (big agri. and the like) notwithstanding

[2] There's nothing preventing a pension fund to go take all of GM's money and put it into a private hedge fund, throwing tons of money at a PE firm, becoming a large stake holder in a real estate investment group, or throw their money into any sort of 'unconventional financial entity'. None of these have public reporting obligations, and their federal obligations are limited to tax entities ,co-operating with FINRA, and (if they are using instruments which fall under the regulatory domain of the SEC, then also) the SEC.

[3] https://en.wikipedia.org/wiki/Export_of_cryptography_from_th...

[4] Both internal and external-- you can do a lot with cost center accounting between arms of an MNC to misrepresent things

The disagreement lies in "I can mathematically apply the trace function, I'm just not quite sure when".

Alternatively, according to the Everett interpretation (to which I subscribe, and to which you seem to as well) quite a few reputed physicists do believe that the universe "splits" (though not in an actual literal sense, there's nothing ripping the universe physically in two) every time an "observation" occurs and there are lots and lots of occurring simultaneously taken from various frames of reference. This actually makes a lot of sense. Let Sean Carroll https://youtu.be/ZacggH9wB7Y convince (the plural) you.

Ethically, were he to enforce the law to the letter with an egalitarian approach, the first person he saw as he got on shift exceeding 65 should have been pulled over. As soon as that ticket was issued, he should have done so for the next person, ad infinitum until he was off-duty.

In actuality, if every single arrested individual stopped one day and decided to exercise his or her constitutional right to see his day in court like that prep school[1] individual did, we'd see what would amount to be a DDoS on the system incapable of handling a caseload of such magnitude. Pragmatically, an ADA will use a very effective method - known colloquially as "charge stacking" - to intimidate the accused into pleaing out a charge he would likely beat[2].

Though I'm totally with you on the classist component - should Michael Jordan's son commit a crime, he'd be far less likely to be convicted than whitey-mc-Kentucky-Meth-Mouth Tim's son due to the same lack of resources afforded to him.

[1] RE: your rape scenario -- Compare this aggressive attorney http://edition.cnn.com/2015/08/28/us/new-hampshire-prep-rape... vs these less fortunate https://en.wikipedia.org/wiki/Central_Park_jogger_case

[2] Even more compelling is the motion for dismissal those defense attorneys could raise in aggregate on 6th amendment grounds. In the state of California, the 45 day cap could be breached fairly easily if all of the defense attorneys concurrently suggested to their clients to enter a not-guilty plea. You'd also see a marked decline in arbitrary traffic stops, as arresting officers would be forced to spend the day in court rather than in speed traps.

Modern Swiss banking privacy is for the DuPont heirs. Their job is to carefully study international tax code, remittance policies of specific countries, volatility of currencies, and a bit of family law within the locales of their clients to minimize divorce losses. Swiss banks exist because they offer a continuity of service. You bank with Julius Baer because your father banked with him. He was at your christening. The overlap between those who crypto-currency user and a private bank client (as anything other than a 'haha, alright let's throw 100k EUR at it and see where it lands' in a speculative fashion) is minimal.

This is a token gesture to the US government saying effectively - hey, fine, Homeland Security, go through any record you like, as long as it's not a Swiss national record. If we were unknowingly complicit in a wire transfer from a Saudi oil account to a Yemeni terror cell, by all means, go for it. Simultaneously, those Carnegie steel heirs are still protected because their name might be listed as "Director" few holding AG's but "sorry, no earning assets there. Any retained earnings go directly to the parent company, we'd love to help you out but it's a Lichtenstein AG. You'll Have to speak to their banking administration." (Oh fun fact, Lichtenstein has more registered companies than it does citizens.) This [0] is how an average run-of-the-mill Swiss Bank constructs their protection. Look at how that institution itself is structured --

    Private Investment Bank Ltd is a fully licensed Bank in
    the Bahamas, established in 1982. It is wholly-owned by
    Banque Cramer & Cie SA, Geneva, Switzerland, which
    is a subsidiary of Norinvest Holding SA, a company
    having family origins in the financial as well as in the
    industrial sector, dating back to 1871." 

Think of banking privacy a lot like the security within your home. An average person with a Kwik-key frontdoor can get bump-keyed or pin-raked in well under a minute[1]. This is basically like opening an auxiliary bank account in a confidant's name to protect against divorce asset issues in a low net-worth scenario (a good attorney will be able to bypass something like that trivially, leading you down forfeiture road).

You throw in a security key like this fancy schmancy German engineered key[2] that's resistant to standard rake-pick techniques, you've upped the entry barrier (haha ugh bad pun intended) from "common thief" to "a professional has likely targeted specifically your house on the Hamptons for that original Matisse you bought at Sotheby's". The banking equivalent of that German key is consulting with UBS or HSBC's private wealth division. If you come from old money, your family has already established all the right safe guards against shirt-loss in divorce, the right non-profit foundations structured, and proper holding companies located in the right tax domiciles (think: Google and 11% in Ireland) structured on any profits yielded.[see: 4]

Sufficient protection for civil suits (i.e., that professional art thief), but your 'lock' won't hold up to the hydraulic battering ram[3] that SWAT uses to break down your door after a ADA goes before a judge alleging you're involved in a terror plot and/or involved in a large drug operation, providing the judge with what seems to be sufficient evidence to sign off on a no-knock warrant. (Though I'd suppose, those battering rams would certainly provide enough 'knock'.) Your HSBC private wealth account might provide you with enough insulation against municipal prosecutors, but certainly not the US DHS at that point.

Here's the thing though - at that point, Swiss banks don't want your money. They'll keep the DuPont and Koch fortunes safe from gold-diggers but they don't want darknet money used to purchase guns, child porn, drugs, or anything even remotely connected to terror. Anything involving proceedings at the first-world federal investigator level (IRS grey-suits notwithstanding, those meetings occur with the run-of-the-mill regularity like the clockwork of a good Patek Philippe [sorry, I couldn't resist]) is a severe impedance for business as usual.

[0] https://www.banquecramer.ch/dam/jcr:7813a747-9c8c-4101-8091-...

[1] The good lock-pickers won't even leave deformation marks on the unit, since they've basically got their hands calibrated to the point where they can feel when their pick engages on a pin.

[2] http://www.theamazingking.com/images/locks/dom-key.jpg

[3] https://en.wikipedia.org/wiki/Door_breaching#/media/File:Trg...

[4] https://www.banquecramer.ch/dam/jcr:7813a747-9c8c-4101-8091-... Here's how a typical private bank structures the assets of their clients to protect against divorce.

If you're the average SV guy who 'daily drives' his Tesla 20 miles from Oakland to his lofted startup where there's 220 to full-charge before you go home, you'll get 2%. Johnny in Kentucky working the coal miles doesn't have that luxury and will certainly enter into 'deep charge' consumption. (600 cycles -> ~75%, with 2nd deriv of batt life w/r/t cycle being negative, i.e., progressively decreasing losses).

Again, not in the field professionally, but these opinions are consistent with my friends who are working at the forefront (albeit, a statistically small sample space, I openly concede !)

I'm all for pure-electric cars but we're still a long way from Joe the miner in Kentucky from being able to drop $1800 on a 2000 Toyota Tundra and having it be able to get him reliably from the jobsite and back. That's not even factoring in the whole capacity-decrease-with-use (and even non-use-- deterioration occurs simply by just storing cells at full capacity for long duration -- of lithium.

Even with the best, most conservative profiles on a battery-module controller for anything lithium based, good luck getting > %50 of cell capacity 5 years down the road of a daily driver [edit: 7]. Anode deterioration (at least, last I seriously researched it for projects requiring portable units for driving larger loads than an average car was ~1.5 years ago) was still a problem even in the lab.[3]

My sister abuses her 2004 Civic coupe to the point where I think she's still running a stock air filter and runs 30-35k between oil changes[4]. This was a run of the mill car she's had since god-knows-how-long and she's still getting 22mpg city [5] ~26 highway on an automatic transmission.

tl;dr -- In terms of total costs :

- capital (purchase) / delivery fee

- operational ($ of petrol for ICE per unit travelled/$ of energy from your power supplier per kw/h), insurance

- maintenance (tire wear, brakes, battery module(s) replacement(s)) over, eh, 5 years from out-of-the-showroom-into-your-garage, I'd be surprised if you saw the electric dollar-for-mile-traveled outperform it's ICE counterpart.[6] I'm far from the forefront of Li, but I do have a few friends in that field (both in academia and in industry) -- even the most optimistic don't see pure-electrics reaching a TCO parity point of an ICE for the consumer in less than 10 years.

--

[1] http://www.greencarreports.com/news/1091436_toyota-gasoline-...

[2] https://www.youtube.com/watch?v=FJXgKY2O4po FreeValve as explained by that really enthusiastic "Engineering Explained" 24 year old automotive engineer, dumbed down to the point where even I can grok it.

[3] Rumor had it, DARPA was using some crazy proprietary stuff that managed to completely nullify dentrification, but if they've managed to accomplish that anodic behavior, there's no way it's going to be released for public usage -- rather, it'll remain hush-hush minus 50 PhD's in metallurgy, and Lockheed drones all of a sudden posting performance numbers +30% from the last revision.

[4] She's not using those long-lasting synthetics that have additives to SeaFoam (yeah, I'm using it as a verb) out carbon build-up on the cylinders and what not, in case you're wondering. Just cheapo 5w/30.

[5] That, albeit was with me driving in 'conservative' mode rather than "hmm let's see the 0-160 on this McLaren".

[6] And I'm 100% sure if you bought a 3 year old variant of an ICE vs a pure-electric, it's no contest -- https://www.edmunds.com/car-buying/drive-a-nearly-new-car-fo... -- ignore the link bait title, it's just about the FMV of cars as a function of time.

[7] In addition to the response I made directly to child-poster, I'd like to concede that it is very possible he's ran 43k miles (i.e., literally at least a thousand cycles, likely closer to mid thousands) with a retained 98%. He makes a very valid point in bringing up the variance of cell capacity deterioration. I'd genuinely love to see some cal'd equipment with your standard dummy load and power analyzer log setup to see the data. (I ain't no fancy electron whiz but I can read me a chart or two.)

The take away is that overall capacity is a function of usage. I could probably get in the lab and simulate 43k miles of load in LabView discharging/recharging every 10 miles while keeping the thermal properties controlled as all heck and see 99.5 capacity retention, but these, again, aren't Joe's driving patterns.

edit 2: @maratd: See my edit 1/[7] (which I presume I was writing while you were drafting your response). I think we're largely in agreement re: usage properties being highly influential. My response to the OC (original child poster) addresses the deep-cycling cooling. As this has turned into a post with 8 endnotes, I think I've crossed the threshold of reasonable discourse.

Allow me to close with a quick remark re: BMC's on your power drill. Anything half decent will have active thermal monitoring specifically because of the reason you stated (much to the chagrin of blue-collar workers everywhere). "Ok, last weld before quittin' time..." paddle trigger actuates, worker expects wire-wheel to start spinning a to clean the slag of iron oxide off the root weld. nothing happens because the thermocouple on the motor armature triggered a lock-out "PC LOAD LETTER WHAT THE HELL DOES THAT MEAN?"

The same thing is true for the fingerboard, the bridge, the strings and other components you interface with (or are interfaced with from things with which you interface) on your violin. As the oils on your finger on the strings/fingerboard gradually acts as basically a mini-abrasive tool, the fingerboard itself will change shape (you can see this easily in the wear patterns). The pressure variance from the string's depression, the shifts in temperature/humidity in the ambient environment (which will stress the wood in all sorts of ways, both in an elastic (temporary) as well as a plastic (permanent) manner), all of these micro-variables add up over time, just like with your car.

I'd imagine if 5 of the best sawyers with 5 of the best luthiers and sound engineers in the world all sat down together with the goal of replicating the Strad's sound with 100% accuracy I'm confident the 'unique' sound of the Strad could be replicated. It might take a lot of time and maybe a few hundred k in equipment (a dozen condenser mics placed strategically at a variety of places within the room and near the instrument itself, some vibration analysis equipment on the equipment, etc), but eventually the sawyer will choose the right wood, the engineer would perform acoustic analysis and tell the luthier "ok, that 43 micron chisel shave you just took just did __, 22 more and we match the F# perfectly on the G string".

I think it's very similar to the whole audiophile thing, where at least some of it is placebo. Those $30 monster cables aren't transmitting with any more signal fidelity than your $5 ones (assuming equal gauge copper, proper ohmic termination, blah blah), and I can sit down and prove it to anyone with $400 worth of spec-ans from the 1980s.

Now the "different" vs "better" -- I love my record player but when audiophiles say 'it sounds better' than the raw mix-down digital masters, in terms of the amount of audible audio content objectively this isn't the case. (Edit: See [4]) It sounds different from what you're used to. You like the 'warmth' that a belt drive record through a 1950s speaker produces. You like the hisses and crackles and pops. It's an emotional connection you have. When I was playing guitar with my crappy band, I'd run my digital content through a Tascam quarter inch tape deck just to get that hiss in the tracks before we sent it down for mastering (we didn't have those fancy plug-ins to do it for us then).

The same thing happened with film.[2] TV is generally shot at 30fps, film at 24fps[3]. When the FPS rate increased with the advent of more modern technology and higher sampling rates, people would claim that cinema didn't look 'cinematic'. The motion blur of the analog film experience we grew up with (well, those of us over 25) is something we mentally associated with the experience of going to the cinema. You saw a bunch of people comment on this with the Hobbit being distributed and shown at theatres at 48 FPS. "It just doesn't look right".

--

[1]http://www.thestrad.com/blind-tested-soloists-unable-to-tell...

[2] I'm sorry I don't have a source on this study, but the phenomenon was pretty widespread. Cinema isn't something I've delved too far into but the cinematographer forums have thousands (literally) of threads on FPS discussion.

[3] That's just historically how things broke down -- 24 was a convenient number to match up with actual seconds due to it's easy factorability, so you could get half a second at 12 frames, a quarter at 6, etc. This made it easy back in the day when post-processing involved actually involving cutting nitrate-based film and sound syncing to the video involved physically matching the snap of the clapper on your audio track to the visual of the actual frame of when the hinge closed.

RE: 30 fps, I'm not sure, but I'd guess it had to do with the fact that we're at 60 hertz in the US so you could interlace half a frame a cycle (ie. trigger on the peak of every AC cycle to update the set of horizontal lines modulo two, so the electron gun presumably only had to hit half the lines within that time constraint) Other parts of the world operate at 50hz, which is why I'd imagine you have the PAL standard and 25 fps in those locales. (Pure speculation though, someone who's knows for sure, jump in.)

[4] Operating under the assumption that that they were pressed from the same master source.. I'm limited only to the records I bought which were from bands with too small of a budget to afford two separate masters. (I.e., That mid-90s emo 200 run EP by Knapsack didn't have two masters, hell on tour they were lucky to have a floor to crash on.) Still, poster beneath me makes a very valid point that should be taken into consideration (Though I'm sure you're already aware of whether or not the band had a separate master DAT/master plate made if you like the band enough to care about the nuanced differences between two masters.)

The problem you end up getting is basically the dual of the 'expression problem'. You get easy type expandability with class augmentation by these semantics, but any client of the type is subsequently bound to deal with any augmentation of the type. (I.e., if in the method defining class 'a', you insert an elseif in between the consequent and the alternative of type Bool, you now just extended the class signature of 'a' to Int32 | String | Bool. Assuming all consumers of 'a' need exhaustive handling, you've opened your class and extended its functionality at the expense of code which had previously satisfied the dependencies to consume 'a' safely.

Crystal still remains a real interesting project to follow. I haven't examined the Crystal type system too carefully, but the concept of Parent+ as virtual types is an interesting concept that is certainly worth exploring.

Microsoft is playing at the integration game. The Surface Hub is a knock-out, they gave it a decent whirl at the Surface Pro 4 (honestly, it's the best far note-taking device on the market IMO simply because the PixelSense display[1] is so stylus friendly - writing on it is nearly like writing on paper), and their very very crummy phone apparently is going to run native Win10 binaries on the new 4-core ARMs so they can leverage all those previous apps and throw them into the ecosystem.

[1]http://www.anandtech.com/show/9767/microsoft-surface-book-20...