I went in to the AT&T store recently to try to fix a problem with my phone. The whole place is done up like a student lounge, with sparsely distributed tables surrounded by cool-looking but uncomfortable chairs. There's no counter, there's nobody obviously in charge. I stood around for a little while, looking around for someone to talk to, but the only visible employee was sitting at a table and talking to someone. I eventually wandered off and did some other shopping.
When I came back a little later, I was able to talk to an employee. He obviously had no technical knowledge at all. He barely listened to my description of my problem, off-handedly suggested that I Google it, and then started trying to sell me a new phone. And trying to pitch me on satellite TV, although I told him over and over again that I wasn't interested.
I started thinking about buying a new phone, and I'm asking him how much it costs. It's hard to get a straight answer, and every figure he quotes includes rebates for buying satellite TV. I finally convince him I don't want that, I just want to buy a phone, maybe with a trade-in allowance from my current phone. Turns out it's not possible at this store to buy a phone in cash. I have no choice other than to sign up for financing.
Nobody was overtly rude, but the whole experience was mostly a waste of time. The one benefit I got from it was the knowledge that there's no point in ever going to the AT&T store.
Yeah, so? Is there a rule that we can't have double standards? Is the point you're trying to make that you have some childish concept of "fairness" that everyone else needs to adhere to? Diane Green is preferentially supporting women because she perceives that men are already preferentially supported in the industry. Is she right? I dunno, I don't have that data. But what's the consequence if she's wrong? That some number of women get VC support who wouldn't without her?
Oh, no. Someone, please protect us from the bad lady giving out money.
I mean, I don't think discrimination on basis of gender is uniformly great, but I do think there are jobs that, in general, men or women are better at than each other. In reality, there are certainly women who can do a 'typically male' job better than some specific men. And vice versa.
Some rich lady has the right to discriminate on the basis of gender when she decides whom to give her money to. The fact that I don't think gender is the best predictor of success isn't important to her. Or, at least, she hasn't consulted me, yet.
I could muster some outrage for a federal grant program discriminating in this way, but I think Diane Greene should do whatever she wants to do. I just won't expect any of her funding.
You must be new here. It's a Slashdot tradition for the editors to re-post stories, even when they're currently on the front page. And it's tradition for us to bitch about them. But enjoy it! Joke about Natalie Portman! Harvest the earlier story for popular comments you can use to karma whore! Imagine a Beowulf cluster of lame iPods, each with less space than a nomad!
I experienced this, in a small way. I was a consulting statistician at the university when a professor from "food sciences" came in and wanted me to "give her the p-value" for her study. I spent some time talking with her about the project that she'd done. It was incredible... tiny sample size, entirely self-selected sample, data based solely on self-evaluation, unaccounted for dependence between responses, likely acquiescence bias, no attempt at control, poorly designed survey instrument, ambiguously written hypotheses... and there was probably more wrong with it that I don't recall off the top of my head.
I told her that maybe she could treat this project as a preliminary data collection and improve the protocols for the real thing, but she was insistent. It turns out she'd already submitted a poster to a conference and had it been accepted and she was leaving the next day and if she didn't have a p-value, then "nobody would take her seriously." I have no idea what she eventually did, but she certainly didn't get a p-value from me.
It just happens sometimes, when you're a statistician, that people come to you and try to get you to bless their study with your magic stats wand.
Talk about a terrible UI... I have family I visit now and then who have Xfinity. Their service is so flaky that they have to reboot their cable box pretty frequently by unplugging it and plugging it back in. It takes something like 3-5 minutes to boot up. What on Earth is going on in there? Once it's booted, any interaction with the UI, including simply changing channels, takes multiple seconds to execute. Why are people willing to put up with this level of input latency just to watch TV? How can whoever designs these things be so terrible at their jobs that this is how poorly their products work?
IANAL, but I've done some research on database copyright law in the EU. So, in the EU, databases (which aren't necessarily data kept in a DBMS, but just collections of data) are not copyrightable except in case of 'sui generis' databases, which are copyrightable 'if they constitute intellectual creation by virtue of the selection or arrangement of their contents.' The conditions a database needs to meet in order to constitute intellectual creation have always been a little unclear to me.
Anyway, whoever wrote the struck-down clause was trying to affirm that these data do not qualify as sui generis, and therefore cannot be copyrighted. But just because the clause was removed doesn't make it obvious that the data do qualify as sui generis and therefore are eligible for copyright. I suspect it kicks the question down the road to some kind of court proceedings. But that's just a guess; it'll be interesting to see what happens.
Years ago, I wrote a spam filter for my email (there already existed better filters than mine, but I wanted to see if I could write one). So I made a Bayesian filter model and fed it a bunch of email, spam and nonspam. Looking at the results of the model, I was interested to find one string that, when present, correctly identified spam without false positives. This string, by itself, was able to filter out about 80% of spam in my inbox.
The string was #FF0000. It turned out that in all the email I'd fed in to the model, I had never received a legitimate email with HTML encoding for red text. At the time, I guess red text was popular in spam (I have no idea if it's still as big). It was an extremely effective filter, but the potential for false positives is pretty clear.
Anyway. Just an example of a computer simplifying a pretty complex problem down to a super-simple answer. Maybe too simple.
I see a lot of partial answers in this thread, but I wanted to list some of the steps in preparing printed documentation of the quality we had in the 80s just to give an idea of the scope of the project. This is long, but not even comprehensive.
You need a technical technical writer. These people are trained and experienced at writing documentation, and they cost a lot of money. If you make very many or complicated products, it might be best to have this person on staff full time. Ideally, they would sit in on meetings between the UI design people and engineers/developers so they understand the product at a basic level.
You need a graphic designer, to create a template for your docs. These people are also trained and experienced, and have the skills to help you create documentation that has a coherent visual hierarchy and style. You can probably just have a consultation with one to create a template you can use for many documents.
You need translators to localize docs. Even shipping to just Canada, you need both English and French documentation. Mostly contractors.
You need editors to review the docs produced by the tech writer. This second set of critical eyes is important to the quality of the document. You might contract this out, especially for other languages.
You need to test the docs. Engineers/devs should review the docs for accuracy, and people outside the dev process should review the docs for ease of use and correctness.
These things would all ideally be done for online documentation, whether HTML- or PDF-based. They're all expensive, which is a big reason we don't have 80s quality of documentation, even on the web. But there are other problems with printing:
You need to print the docs. This happens before the product is finished and on the assembly line. Here is an opportunity for errors in the docs, because last-minute design changes make the docs obsolete.
The docs go in the box, adding somewhat to packaging and shipping costs.
With each firmware patch and software update, the docs grow increasingly out of sync with the product. Consumers don't like docs that don't match their product.
Since docs go out of date, and consumers lose docs and want replacements, the easiest way to provide users with the latest docs is on the web. Why bother printing in the first place, when the web is the superior means of distribution? No printing cost, no handling or shipping cost, no users angry because of obsolete docs.
Finally, there are cynical reasons for not providing high-quality documentation. An obvious one is support contracts. Whether by phone/email or on-site visit, the company can charge a lot of money to help you use their poorly documented product.
Or, just leave the users to their own devices. If your product is popular, the users will all jump on somewhere like Reddit and build their own documentation, some of them obsessively. Voila, free docs at zero cost.
My ~5 year old HP laptop was slowing down and had frequent glitches with the Wifi disconnecting. I bought it a new SSD and did a fresh install of Windows 7. The SSD much improved speed, as did wiping out whatever garbage had built up in the registry. My Wifi problem is gone. It's almost like a new (not the newest, but newer) machine.
Back in the Win 9x and XP days, I'd reinstall Windows every year or so just to clear out cruft and keep the performance up. Windows 7 was a big improvement, but it looks like it still benefits from the occasional wipe every few years, under my usage habits.
I feel like it would be pretty easy to identify some statistical characteristics of this type of spam. For one thing, it's really long compared to a real comment. Basic token frequency analysis would probably show it to be clearly different from typical real comments.
One of the admins could pick out a bunch of obvious spam like this, GNAA, the weird shitposts I've seen recently describing the events of someone's day that looks Markov generated. There's plenty of volume of it, and it should be pretty easy to find by reviewing -1 comments. Gather it up (with legitimate posts for comparison) into a corpus and feed it into a Bayesian filter and use the results to power an automoderator. At first, the automod mods spam to -1. Let users metamod the automod to weed out false positives (but with the spam content being so significantly different from real content, I feel like the classification should be pretty good).
Let that go on for a while until they're satisfied with automod's performance, and then set the automod to mod spam down to -2. Don't allow normal users to mod to -2. Don't allow -2 posts to be further moderated. Add -2 to the message filtering controls, with -2 messages being hidden by default. If we collectively agreed to ignore -2 posts, then we have created a system that effectively shitcans spam, even though it's still there, by making it invisible. Spammers can continue to post crazy racism, and it will just flow into a gutter that we made for it. Users can continue to metamod automod, so we could continue to catch false positives.
One weakness is that modding to -2 makes it obvious that the automod caught a post, and a clever spammer could use that fact to refine their post to avoid automod. But we throttle anonymous posts, so that might be hard to do. I guess you could come up with some more elaborate shadowban system, so it appeared to the spam poster that their message was up, but it sounds like a lot more work for not much payoff.
How much are you willing to pay to be able to play games on ultra settings on computers that don't meet the minimum system requirements for those games? How much are you willing to pay to play a AAA game on a Mac that hasn't been released for Mac. PC games on Chromebook?
It's a good point, and one that I've played around a little bit with Steam streaming. I have my powerful desktop computer in my office, but I can install a game on it and then go out and play it on my laptop in the living room (even on my MacBook Air, I suppose, although I stream to a laptop with a larger screen). And the performance is pretty good. Not always great; I see some I/O latency sometimes. But then I've done nothing to try to optimize my streaming experience. Plus, they talk about streaming games being run on machines with huge graphics arrays that allow for rendering far beyond the abilities of my desktop.
I see the writing on the wall. Game publishers would be crazy to go any way other than streaming, because it allows them to have total control over the binaries for their game. No more piracy, no more cheating, no more user modding, plus they get to charge in perpetuity for every game.
I mean, I was reluctant to get on board with Steam, and one day I bought a boxed game at the store and brought it home and inside the box was a Steam code. At that point, I gave up and started buying stuff in Steam, and today I have a library of 170 games there.
Have contingency plans for every option they can choose... This principle has been working for Choose Your Own Adventure books
In my experience with Choose Your Own Adventure books, the contingency for every option I chose was a gruesome death. So, following that model, they could prepare for many contingencies with just one plan!
On the cynical hand, it seems to me that game streaming is just another (perhaps the ultimate) way for a company to demand that I pay monthly rather than just buying the software I want.
But on the other hand, it seems like there could be real value to it. With streaming, I could play any game on Linux or elsewhere, while the developers only have to support one OS. I don't have to worry about installing games anymore, or patching them. I never have to worry about driver incompatibility (or maybe not never, but less).
Despite whatever positives it might have, I'm reluctant to get on board. I don't want to have to be online to play. I do want to be able to keep playing after the streaming service abandons a game I like. I don't want to pay monthly, even if I get access to a larger library of games; I'd rather buy individual titles I want.
I guess I just don't see it as an advantage until the payoff to me becomes much greater. Right now, the benefits mostly seem to be in favor of the businesses.
I mean, as long as it's cheaper to suffer IT outages than it is to fund IT, clearly they'll keep offshoring. It doesn't look like a sustainable model to me. I understand they're going to do the more profitable thing, but when this tech debt catches up to them in the form of a catastrophic outage, will they be able to withstand the fallout? I'm sure they have risk management people who model this out and say the answer is yes, but those guys aren't infallible.
Delta reported to investors that their 5-hour outage in 2016 cost $150 million. Whatever the numbers are, I guess they believe they save money pay not having in-house IT. It seems risky, though, if 5 hours costs you that much. What happens when your offshored contractors cause a 10 hour outage? I doubt that the cost of the outage scales linearly with time... I'd guess it goes up faster than that.
How cute that you think airlines actually OWN the airplanes
Condescending tone aside, it's not that important what the financing mechanism is for the aircraft. Airlines have a lot of money tied up in inventory. Maybe most maintenance is done by foreign contractors, but it's also done following recommendations by the manufacturer and regulations by the FAA. Plus, it's real bad business when a passenger airplane falls out of the sky because the operator cheaped out on maintenance. Given the rarity of plane crashes due to poor maintenance (Wikipedia lists 8 or so in the US since the FAA has existed), it seems that the airlines are doing okay there.
the airline business is one of managing debt and cash flow
And my point is, while your planes are all grounded because you offshored a mission-critical part of your enterprise to a bunch of guys who are just about half-competent and have just about half a shit to give about your business, your cash flow is all going one way. In 2016, Delta IT systems had a 5 hour outage that cost $155 million. They've had fleet-grounding IT failures every year since then. As their infrastructure continues to decay under outsourcing, I don't see these outages getting shorter or less frequent.
I'm not surprised, although I keep thinking that the massive amounts of money Delta loses due to IT failures would be enough incentive to bring their IT infrastructure back under their control. Since 2016, Delta's CIO, Rahul Samant, has been working to move IT employees out of Delta into Indian IT consultancies (see http://www.fox9.com/news/delta...).
It seems to me that airlines basically have four things of real value: aircraft, ground crews to maintain the aircraft, flight crews to operate the aircraft, and an IT system that allows them to schedule and dispatch the aircraft. When an airline talks about its "core competencies," I feel like their IT system should definitely be one, because if it fails, the planes don't fly. That seems like enough reason to not offshore that part of the business.
I'll take that as a serious question. The last time I tried to really use LibreOffice, I went all the way and installed only Linux on my desktop, and LibreOffice came with. I had never used it much before, but I had no particular bias against it. I created a PowerPoint presentation on my Windows laptop, and then transferred it to my desktop to continue working on it. When I opened it in LibreOffice, it fucked my fonts, formatting, placement, and just about everything else. My text was the only thing that was properly loaded. My slides were useless as rendered by LibreOffice. I spent some time trying to fix them, but eventually decided to give up and finish my slides on my Windows laptop. I stopped using LibreOffice because it was effectively incompatible with MS Office, and I actually need to be able to interact with MS Office documents. I say this as a supporter and sometimes evangelist of open source software.
LibreOffice is a good program, but it is *not* fully interoperable with MS Office. In my experience, LibreOffice cannot be realistically used to collaborate with colleagues who are using MS Office. I'm sure it works great standalone or in a group environment where everyone is using it.
I am not a Microsoft fanboy. I'd rather use Linux as my primary OS. I admire the concept of LibreOffice. But MS Office, in my experience, is the only software that truly interoperates with MS Office. That doesn't make it good, but it does make it the industry standard.
I work pretty frequently with other people who are on Office 365, and I haven't experienced any incompatibilities in Word, Excel, or PowerPoint. These are mostly pretty basic business documents, so it's possible that there are problems I'm just not running into because none of us are using incompatible features.
If you can, check the packaging to see how many computers you can install it on. My version allows me to install it on ~3 computers, but I know they changed the licensing at some point to only allow 1 installation. (I think I saw an article the other day where they're talking about changing it back again.)
Meanwhile, I've been using my locally installed copy of Office 2010 without any trouble at all. I'm even blissfully unaware of all the new and innovative features of Office in The Cloud.
It's a perfect recipe for gaming the ratings system with quid pro quo. Drivers are punished for ratings lower than 4.6, passengers are punished for ratings lower than 4. Both parties are now incentivised to give each other 5 star ratings, and both have leverage against each other to prevent lower ratings.
The only thing this can possibly accomplish is to further devalue the ratings system, itself. I guess it will make middle managers happy with the metrics to see that 100% of drivers fall within the top 10% of drivers, and 100% of passengers fall within the top 20% of passengers. Those are really great numbers.
I just assume that any online (cloud based or not) OCR or fax bridge site is going to store a copy of my document in an insecure way. I assume that employees of the service will have access to view my document. I haven't thought too much about them exposing my documents to the public, but it's not a huge step from what I already assumed about them. Anyway, the result is that I don't send anything sensitive or with information I wouldn't want publicly known through online OCR or fax. Because it would be crazy to upload my private sensitive documents to randos on the Internet and assume that they'll never be seen.
Slashdot Mirror
I went in to the AT&T store recently to try to fix a problem with my phone. The whole place is done up like a student lounge, with sparsely distributed tables surrounded by cool-looking but uncomfortable chairs. There's no counter, there's nobody obviously in charge. I stood around for a little while, looking around for someone to talk to, but the only visible employee was sitting at a table and talking to someone. I eventually wandered off and did some other shopping.
When I came back a little later, I was able to talk to an employee. He obviously had no technical knowledge at all. He barely listened to my description of my problem, off-handedly suggested that I Google it, and then started trying to sell me a new phone. And trying to pitch me on satellite TV, although I told him over and over again that I wasn't interested.
I started thinking about buying a new phone, and I'm asking him how much it costs. It's hard to get a straight answer, and every figure he quotes includes rebates for buying satellite TV. I finally convince him I don't want that, I just want to buy a phone, maybe with a trade-in allowance from my current phone. Turns out it's not possible at this store to buy a phone in cash. I have no choice other than to sign up for financing.
Nobody was overtly rude, but the whole experience was mostly a waste of time. The one benefit I got from it was the knowledge that there's no point in ever going to the AT&T store.
Yeah, so? Is there a rule that we can't have double standards? Is the point you're trying to make that you have some childish concept of "fairness" that everyone else needs to adhere to? Diane Green is preferentially supporting women because she perceives that men are already preferentially supported in the industry. Is she right? I dunno, I don't have that data. But what's the consequence if she's wrong? That some number of women get VC support who wouldn't without her?
Oh, no. Someone, please protect us from the bad lady giving out money.
I mean, I don't think discrimination on basis of gender is uniformly great, but I do think there are jobs that, in general, men or women are better at than each other. In reality, there are certainly women who can do a 'typically male' job better than some specific men. And vice versa.
Some rich lady has the right to discriminate on the basis of gender when she decides whom to give her money to. The fact that I don't think gender is the best predictor of success isn't important to her. Or, at least, she hasn't consulted me, yet.
I could muster some outrage for a federal grant program discriminating in this way, but I think Diane Greene should do whatever she wants to do. I just won't expect any of her funding.
I mean... yeah? That's okay. I'm sure there are venture capitalists out there who make decisions that way.
Or were you trying to use your incredibly strained butthurt here to provoke some kind of outrage?
Same for me. My phone works fine in the microwave with the door shut. However, wrapping it in tin foil did block the signal.
You must be new here. It's a Slashdot tradition for the editors to re-post stories, even when they're currently on the front page. And it's tradition for us to bitch about them. But enjoy it! Joke about Natalie Portman! Harvest the earlier story for popular comments you can use to karma whore! Imagine a Beowulf cluster of lame iPods, each with less space than a nomad!
I experienced this, in a small way. I was a consulting statistician at the university when a professor from "food sciences" came in and wanted me to "give her the p-value" for her study. I spent some time talking with her about the project that she'd done. It was incredible... tiny sample size, entirely self-selected sample, data based solely on self-evaluation, unaccounted for dependence between responses, likely acquiescence bias, no attempt at control, poorly designed survey instrument, ambiguously written hypotheses... and there was probably more wrong with it that I don't recall off the top of my head.
I told her that maybe she could treat this project as a preliminary data collection and improve the protocols for the real thing, but she was insistent. It turns out she'd already submitted a poster to a conference and had it been accepted and she was leaving the next day and if she didn't have a p-value, then "nobody would take her seriously." I have no idea what she eventually did, but she certainly didn't get a p-value from me.
It just happens sometimes, when you're a statistician, that people come to you and try to get you to bless their study with your magic stats wand.
Talk about a terrible UI... I have family I visit now and then who have Xfinity. Their service is so flaky that they have to reboot their cable box pretty frequently by unplugging it and plugging it back in. It takes something like 3-5 minutes to boot up. What on Earth is going on in there? Once it's booted, any interaction with the UI, including simply changing channels, takes multiple seconds to execute. Why are people willing to put up with this level of input latency just to watch TV? How can whoever designs these things be so terrible at their jobs that this is how poorly their products work?
IANAL, but I've done some research on database copyright law in the EU. So, in the EU, databases (which aren't necessarily data kept in a DBMS, but just collections of data) are not copyrightable except in case of 'sui generis' databases, which are copyrightable 'if they constitute intellectual creation by virtue of the selection or arrangement of their contents.' The conditions a database needs to meet in order to constitute intellectual creation have always been a little unclear to me.
Anyway, whoever wrote the struck-down clause was trying to affirm that these data do not qualify as sui generis, and therefore cannot be copyrighted. But just because the clause was removed doesn't make it obvious that the data do qualify as sui generis and therefore are eligible for copyright. I suspect it kicks the question down the road to some kind of court proceedings. But that's just a guess; it'll be interesting to see what happens.
Years ago, I wrote a spam filter for my email (there already existed better filters than mine, but I wanted to see if I could write one). So I made a Bayesian filter model and fed it a bunch of email, spam and nonspam. Looking at the results of the model, I was interested to find one string that, when present, correctly identified spam without false positives. This string, by itself, was able to filter out about 80% of spam in my inbox.
The string was #FF0000. It turned out that in all the email I'd fed in to the model, I had never received a legitimate email with HTML encoding for red text. At the time, I guess red text was popular in spam (I have no idea if it's still as big). It was an extremely effective filter, but the potential for false positives is pretty clear.
Anyway. Just an example of a computer simplifying a pretty complex problem down to a super-simple answer. Maybe too simple.
I see a lot of partial answers in this thread, but I wanted to list some of the steps in preparing printed documentation of the quality we had in the 80s just to give an idea of the scope of the project. This is long, but not even comprehensive.
These things would all ideally be done for online documentation, whether HTML- or PDF-based. They're all expensive, which is a big reason we don't have 80s quality of documentation, even on the web. But there are other problems with printing:
Finally, there are cynical reasons for not providing high-quality documentation. An obvious one is support contracts. Whether by phone/email or on-site visit, the company can charge a lot of money to help you use their poorly documented product.
Or, just leave the users to their own devices. If your product is popular, the users will all jump on somewhere like Reddit and build their own documentation, some of them obsessively. Voila, free docs at zero cost.
My ~5 year old HP laptop was slowing down and had frequent glitches with the Wifi disconnecting. I bought it a new SSD and did a fresh install of Windows 7. The SSD much improved speed, as did wiping out whatever garbage had built up in the registry. My Wifi problem is gone. It's almost like a new (not the newest, but newer) machine.
Back in the Win 9x and XP days, I'd reinstall Windows every year or so just to clear out cruft and keep the performance up. Windows 7 was a big improvement, but it looks like it still benefits from the occasional wipe every few years, under my usage habits.
I feel like it would be pretty easy to identify some statistical characteristics of this type of spam. For one thing, it's really long compared to a real comment. Basic token frequency analysis would probably show it to be clearly different from typical real comments.
One of the admins could pick out a bunch of obvious spam like this, GNAA, the weird shitposts I've seen recently describing the events of someone's day that looks Markov generated. There's plenty of volume of it, and it should be pretty easy to find by reviewing -1 comments. Gather it up (with legitimate posts for comparison) into a corpus and feed it into a Bayesian filter and use the results to power an automoderator. At first, the automod mods spam to -1. Let users metamod the automod to weed out false positives (but with the spam content being so significantly different from real content, I feel like the classification should be pretty good).
Let that go on for a while until they're satisfied with automod's performance, and then set the automod to mod spam down to -2. Don't allow normal users to mod to -2. Don't allow -2 posts to be further moderated. Add -2 to the message filtering controls, with -2 messages being hidden by default. If we collectively agreed to ignore -2 posts, then we have created a system that effectively shitcans spam, even though it's still there, by making it invisible. Spammers can continue to post crazy racism, and it will just flow into a gutter that we made for it. Users can continue to metamod automod, so we could continue to catch false positives.
One weakness is that modding to -2 makes it obvious that the automod caught a post, and a clever spammer could use that fact to refine their post to avoid automod. But we throttle anonymous posts, so that might be hard to do. I guess you could come up with some more elaborate shadowban system, so it appeared to the spam poster that their message was up, but it sounds like a lot more work for not much payoff.
It's a good point, and one that I've played around a little bit with Steam streaming. I have my powerful desktop computer in my office, but I can install a game on it and then go out and play it on my laptop in the living room (even on my MacBook Air, I suppose, although I stream to a laptop with a larger screen). And the performance is pretty good. Not always great; I see some I/O latency sometimes. But then I've done nothing to try to optimize my streaming experience. Plus, they talk about streaming games being run on machines with huge graphics arrays that allow for rendering far beyond the abilities of my desktop.
I see the writing on the wall. Game publishers would be crazy to go any way other than streaming, because it allows them to have total control over the binaries for their game. No more piracy, no more cheating, no more user modding, plus they get to charge in perpetuity for every game.
I mean, I was reluctant to get on board with Steam, and one day I bought a boxed game at the store and brought it home and inside the box was a Steam code. At that point, I gave up and started buying stuff in Steam, and today I have a library of 170 games there.
In my experience with Choose Your Own Adventure books, the contingency for every option I chose was a gruesome death. So, following that model, they could prepare for many contingencies with just one plan!
On the cynical hand, it seems to me that game streaming is just another (perhaps the ultimate) way for a company to demand that I pay monthly rather than just buying the software I want.
But on the other hand, it seems like there could be real value to it. With streaming, I could play any game on Linux or elsewhere, while the developers only have to support one OS. I don't have to worry about installing games anymore, or patching them. I never have to worry about driver incompatibility (or maybe not never, but less).
Despite whatever positives it might have, I'm reluctant to get on board. I don't want to have to be online to play. I do want to be able to keep playing after the streaming service abandons a game I like. I don't want to pay monthly, even if I get access to a larger library of games; I'd rather buy individual titles I want.
I guess I just don't see it as an advantage until the payoff to me becomes much greater. Right now, the benefits mostly seem to be in favor of the businesses.
I mean, as long as it's cheaper to suffer IT outages than it is to fund IT, clearly they'll keep offshoring. It doesn't look like a sustainable model to me. I understand they're going to do the more profitable thing, but when this tech debt catches up to them in the form of a catastrophic outage, will they be able to withstand the fallout? I'm sure they have risk management people who model this out and say the answer is yes, but those guys aren't infallible.
Delta reported to investors that their 5-hour outage in 2016 cost $150 million. Whatever the numbers are, I guess they believe they save money pay not having in-house IT. It seems risky, though, if 5 hours costs you that much. What happens when your offshored contractors cause a 10 hour outage? I doubt that the cost of the outage scales linearly with time... I'd guess it goes up faster than that.
Condescending tone aside, it's not that important what the financing mechanism is for the aircraft. Airlines have a lot of money tied up in inventory. Maybe most maintenance is done by foreign contractors, but it's also done following recommendations by the manufacturer and regulations by the FAA. Plus, it's real bad business when a passenger airplane falls out of the sky because the operator cheaped out on maintenance. Given the rarity of plane crashes due to poor maintenance (Wikipedia lists 8 or so in the US since the FAA has existed), it seems that the airlines are doing okay there.
And my point is, while your planes are all grounded because you offshored a mission-critical part of your enterprise to a bunch of guys who are just about half-competent and have just about half a shit to give about your business, your cash flow is all going one way. In 2016, Delta IT systems had a 5 hour outage that cost $155 million. They've had fleet-grounding IT failures every year since then. As their infrastructure continues to decay under outsourcing, I don't see these outages getting shorter or less frequent.
I'm not surprised, although I keep thinking that the massive amounts of money Delta loses due to IT failures would be enough incentive to bring their IT infrastructure back under their control. Since 2016, Delta's CIO, Rahul Samant, has been working to move IT employees out of Delta into Indian IT consultancies (see http://www.fox9.com/news/delta...).
It seems to me that airlines basically have four things of real value: aircraft, ground crews to maintain the aircraft, flight crews to operate the aircraft, and an IT system that allows them to schedule and dispatch the aircraft. When an airline talks about its "core competencies," I feel like their IT system should definitely be one, because if it fails, the planes don't fly. That seems like enough reason to not offshore that part of the business.
I'll take that as a serious question. The last time I tried to really use LibreOffice, I went all the way and installed only Linux on my desktop, and LibreOffice came with. I had never used it much before, but I had no particular bias against it. I created a PowerPoint presentation on my Windows laptop, and then transferred it to my desktop to continue working on it. When I opened it in LibreOffice, it fucked my fonts, formatting, placement, and just about everything else. My text was the only thing that was properly loaded. My slides were useless as rendered by LibreOffice. I spent some time trying to fix them, but eventually decided to give up and finish my slides on my Windows laptop. I stopped using LibreOffice because it was effectively incompatible with MS Office, and I actually need to be able to interact with MS Office documents. I say this as a supporter and sometimes evangelist of open source software.
LibreOffice is a good program, but it is *not* fully interoperable with MS Office. In my experience, LibreOffice cannot be realistically used to collaborate with colleagues who are using MS Office. I'm sure it works great standalone or in a group environment where everyone is using it.
I am not a Microsoft fanboy. I'd rather use Linux as my primary OS. I admire the concept of LibreOffice. But MS Office, in my experience, is the only software that truly interoperates with MS Office. That doesn't make it good, but it does make it the industry standard.
I work pretty frequently with other people who are on Office 365, and I haven't experienced any incompatibilities in Word, Excel, or PowerPoint. These are mostly pretty basic business documents, so it's possible that there are problems I'm just not running into because none of us are using incompatible features.
If you can, check the packaging to see how many computers you can install it on. My version allows me to install it on ~3 computers, but I know they changed the licensing at some point to only allow 1 installation. (I think I saw an article the other day where they're talking about changing it back again.)
Meanwhile, I've been using my locally installed copy of Office 2010 without any trouble at all. I'm even blissfully unaware of all the new and innovative features of Office in The Cloud.
It's a perfect recipe for gaming the ratings system with quid pro quo. Drivers are punished for ratings lower than 4.6, passengers are punished for ratings lower than 4. Both parties are now incentivised to give each other 5 star ratings, and both have leverage against each other to prevent lower ratings.
The only thing this can possibly accomplish is to further devalue the ratings system, itself. I guess it will make middle managers happy with the metrics to see that 100% of drivers fall within the top 10% of drivers, and 100% of passengers fall within the top 20% of passengers. Those are really great numbers.
I just assume that any online (cloud based or not) OCR or fax bridge site is going to store a copy of my document in an insecure way. I assume that employees of the service will have access to view my document. I haven't thought too much about them exposing my documents to the public, but it's not a huge step from what I already assumed about them. Anyway, the result is that I don't send anything sensitive or with information I wouldn't want publicly known through online OCR or fax. Because it would be crazy to upload my private sensitive documents to randos on the Internet and assume that they'll never be seen.