a) Was Erlang the right choice b) Did management handle transitioning the team to Erlang the right way.
(b) could be "no" and (a) could be yes. Your argument before was that Erlang wasn't a good language. No one is arguing that FP doesn't take unlearning a bunch of ways of thinking that are typical in other languages. Also some pure learning time where you see concepts on simple examples and not in the context of real life. Redoing programming 101 type stuff.
I don't think so. FP is good when you need to do complex things with simple state issues. If you need to do complex things with state and simple things to the data FP are just bad imperative languages.
FP has already taken over the world. Have you looked at what's happening with compilers at this point. Your imperative languages are more and more interpreters running against functional engines (though running against very low level imperative code).
A good example of where functional can be very useful from your examples is: writing the database engine, writing the CD-decoder, writing the network stack, etc...
You can write beautiful, elegant, purely functional code, as long as it doesn't have to touch a storage system, a network, or a user. But, hey, other than that, it's great!
You write beautiful elegant functional code for the algorithms and then messy imperative code that invokes those algorithms for stateful operations. If you do it in the functional language itself Monads work well. Otherwise you just invoke the functional code as a library.
CGN is a fantasy the carriers are clear about that. It is popular on/. with the "can't do" crowd that's taken over IT but not with a single carrier I've ever met. Besides ARIN hates the idea and going to something without regulatory approval is dangerous.
I don't follow your comment about paying though. Even if carriers offered CGN they could drop it at a later date when as the contracts expired and force the move that way and/or just keep raising fees for IPv4. I suspect the fees are what they are going to do.
Why would carriers push other manufacturers subsidies up that would cause then to spend more money.?
For the same reason they introduce subsidies in the first place.
1) The quality of the handset has a huge impact on the level of custom enjoyment of their cellular solution and thus their usage. 2) Customers however are cheap with handsets. 3) Customers are willing to pay more in monthly fees
So it makes sense for carriers to essentially lend customers money for the handsets as part of their monthly fees to get their total spend up. For every $1 in handset subsidies they get $3 in revenue which means $2 in revenue for the carrier related business.
Encourage contract customers to buy cheaper Android phones so they could spend less on subsidies.
That article was during 2012. During that year the subsidies on Android were $250-300 while the iPhone subsidy was $420. Of course Verizon wanted customers to buy an expensive dataplan and then kick back $8 / mo extra. But despite stores being mildly pro-Android 70% of their contract base still went iPhone when they could.
Move away from a subsidy model and offer 0% financing on the phone so that the customer pays for the phone and plan separately.
Note the May 2012 date. They went the other way. They increased subsidies. The article was wrong.
"That word... I don't think it means what you think it means..."
Let me clarify. Their addresses are going to be virtualized relative to the IPv6 space. The are going to look to the IPv6 customers as if they were on Verizon / AT&T... local routable subnet while in reality they are passing through a gateway. I'm not saying the servers are virtualized but the route will be.
Carriers have been "ready" for years--nothing whatsoever is stopping Comcast, AT&T and everybody else from flipping 100% of their users to IPv6 tomorrow,
Not they aren't ready. I've attended conferences on this. Verizon is getting close to ready. Comcast is done and it was 6 years of agony. They still have lots of stuff in their consumer / small business setups that makes v4 assumptions.
But there's no content to access via IPv6... So what's the rush?
The rush is to be able to reclaim that valuable v4 address space to sell it at a premium to business.
10.6 wasn't a free upgrade. $20 is cheap it was during the time when Apple was lowering prices. They aren't defrauding anyone the prices on the upgrades are made public and haven't changed.
I have to say thought that is $20 bothers you, you aren't a good fit for Apple. You'll get hit far worse on other things. You will be much happier in the skinflint world of Linux or Windows.
OpenCL is a huge speed up and OpenGL 4.1 was too. If you go back Core Audio, Core Video, and Core Animation where huge speed ups. For a desktop (i.e. audio video...) I think there is a little doubt that OS X is way faster on good hardware than Linux.
I'm not sure that OS X, in the hands of an average user, is really that much better in terms of longevity.
We have pretty good data and the answer is absolutely yes it is. The overwhelming majority of OSX users never need to do a full OS reinstall. They move config from computer to computer. I'm not an average user but haven't done a clean install on my main system since OS 10.1.
I'm going to agree with damn_registrars. Unless you are really young there is no such thing as a decent new bike for $200. The decent parts cost way more than that. You would be far better off going with a used bike and getting it cleaned up for that money.
That's not true. It was true in 2012 but not in 2013 or 2014. The carriers have been trying to prevent the formation of an Apple monopoly in the USA and pushed Android subsidies up to Apple levels. Apple is currently negotiating for yet another increase for the iPhone 6 so this may happen 4Q2014 but for about 18 mo it wasn't the case.
Once home / small business switch over the content providers are going to be virtualized. Which means that service will stop working, geolocation being the first to go. They'll lose the ability to meaningfully regulate traffic (everything is coming from West Virginia). It is fairly east to switch most websites over. Most consumer content will switch with a few years of the carriers being ready.
No they won't do that. Carriers have been pretty clear they aren't implementing carrier grade NAT and supporting it. ARIN has been hostile to them making use of carrier grade NAT. It isn't happening.
This is a problem. I used to support OpenOffice, until Sun was bought by Oracle. Then I recommended LibreOffice for about as long as it took Oracle to give the package to Apache Foundation
Oracle gave OpenOffice to Apache not Libre Office. Nothing changed about Libre Office. If you could recommend it before you can recommend it now.
Not having a clear winner, nor a clear advantage in operations (moral/functional) makes it the "use either" that will assure neither succeeds.
I don't see any reason that's true. They are mostly the same. That's like saying Microsoft Office 2013 and Office 365 will make sure that neither is successful.
If you don't need document management then SharePoint won't do anything for you. It is entirely possible that a wiki would be a better choice. At this point sing you already have it maybe consolidate down to just a few SharePoint sites like: Engineering, HR, Accounting... That way you can use SharePoint for areas where people are unfamiliar with who to talk to. So for example if you don't know who in Accounting handles foreign checks and what the policy is for foreign purchase SharePoint works for you, but inside engineering you use something else.
First off, they emphasize these ridiculous social networking style features. "Connecting to people". As if you don't know who the members of your team are.
Well Yammer is a new feature. People in large companies often don't know the other people in their teams. You may not have worked in large enough organizations where this is a problem.
. If you are such an MS house that you have Sharepoint, then you already use Outlook as your primary written communication tool - no one is going to switch over to Sharepoint for collaboration unless you take away Outlook, or unless MS spends some time actually integrating Outlook with Sharepoint in more than a token way.
Email is a 1-1 or small-small collaboration system. Document management is many-many.
I still don't understand how Excel/Word/Powerpoint losing data when saving to a Sharepoint site is not what you would consider a major problem with the client software, but a configuration problem on the Sharepoint side. A failure to upload should trigger some kind of fallback to local storage.
Because that's not a SharePoint problem it is a SharePoint as you have it configured problem.
IMHO, you need a dedicated librarian on each team if you want to use Sharepoint (or similar metadata based tools).
I don't know about dedicated but yes you need at least one person in each document creation group who sees SharePoint as one of their core job functions. The librarian is more at the level of integrating across the company. It sounds to me in your description like your company might be too small to have the sorts of problems that SharePoint is designed to fix.
And God, the lock-in. It's bad enough that we have these legacy wikis running - now in emulated hardware purgatory. But at least they are running in a VPN somewhere in purgatory, requiring no admin time and not requiring any license fee. Our little failed experiments can wither peacefully and cheaply - only accessed when old data is needed for reference. But this Sharepoint stuff, if they decide to stop paying the fee it is gone. I'm sure you can export it in some manner, but what a mess.
No argument there. SharePoint is a serious lock in. The companies like yours who are using it wrong have much less of a problem than the companies who are using it right. It will cost those companies a fortune to get their data out. Most likely they will be on Share Point forever.
OK, would it be fair to say that Sharepoint sucks in roles that lots of people seem to expect it to be useful for?
Yes that's fair. Though IMHO most companies and individuals don't really use any of their software right. So for example the #1 thing people do with Excel is keep lists. There is far better outlining software than Excel which is much better for lists. But...
I'm hardly an expert, but I kept getting the impression that it was a bad tool intended for general use.
I think most companies (over 100 employees) need a document library. Most companies write a large percentage of their documents in Office. I think Share Point is a reasonable choice for a document library for most companies. That's the general use. Then you throw in project resources like team calendars, team wikis... I think it is rather good and less hassle than most of them. But that's different than saying no hassle.
Taking the example of a file share. What most people do with a file share is better done in SharePoint than NFS. Which is not to say Share Point makes a good no frills file share but the changes that Share Point induces are the right kinds of changes for most companies and will benefit them.
Similarly on things like project calendar sharing.
I wouldn't say SharePoint sucks. It sucks when you don't use it as intended. The GP was using it as a file share. SharePoint is very good as a light document management solution with MS Office integration.
As far as the rest, I said the comparison is to Documentum. Documentum generally recommends an implementation team. It is not uncommon that Documentum has a permanent IT support department (i.e. multiple people). Share Point can get by with less than 1. As for metadata and using it its way. Yes. That's standard for a document management solution.
____
Imagine if someone was talking about SQLServer as a replacement for Oracle and we were discussing the lower cost and lower administrative hassle relative to Oracle. And you kept coming back with a comparison to using text files and how much easier and less of a problem text files were. SQL Server did a great job of taking Oracle administration (a skilled profession) and making it something that a moderately skilled admin could do part time on the side. Similarly SharePoint gets you 80% of Documentum for 20% of the cost.
1000 users on Documentum fully configured is probably just under $1m. 1000 uses on Sharepoint fully configured is about $200k
OTOH 1000 users on a fileserver fully configured is going to be $20k.
That's where you need to make the comparison. For smaller business, SharePoint online you can be at $7 / user / mo for a small business. There is nothing remotely like that for the higher end suites. But heck I can get 50g file share in the cloud for free.
Slashdot Mirror
There are two issues here:
a) Was Erlang the right choice
b) Did management handle transitioning the team to Erlang the right way.
(b) could be "no" and (a) could be yes. Your argument before was that Erlang wasn't a good language. No one is arguing that FP doesn't take unlearning a bunch of ways of thinking that are typical in other languages. Also some pure learning time where you see concepts on simple examples and not in the context of real life. Redoing programming 101 type stuff.
A0 = array:new(10).
10 = array:size(A0).
Will give you a fixed sized array.
____
If you need a mutable array datatype you just create a monad and do it in there. Which incidentally has been done for Erlang as opensource libraries.
I don't think so. FP is good when you need to do complex things with simple state issues. If you need to do complex things with state and simple things to the data FP are just bad imperative languages.
FP has already taken over the world. Have you looked at what's happening with compilers at this point. Your imperative languages are more and more interpreters running against functional engines (though running against very low level imperative code).
Pure functional programming means state is isolated not that it doesn't exist.
A good example of where functional can be very useful from your examples is: writing the database engine, writing the CD-decoder, writing the network stack, etc...
You write beautiful elegant functional code for the algorithms and then messy imperative code that invokes those algorithms for stateful operations. If you do it in the functional language itself Monads work well. Otherwise you just invoke the functional code as a library.
CGN is a fantasy the carriers are clear about that. It is popular on /. with the "can't do" crowd that's taken over IT but not with a single carrier I've ever met. Besides ARIN hates the idea and going to something without regulatory approval is dangerous.
I don't follow your comment about paying though. Even if carriers offered CGN they could drop it at a later date when as the contracts expired and force the move that way and/or just keep raising fees for IPv4. I suspect the fees are what they are going to do.
For the same reason they introduce subsidies in the first place.
1) The quality of the handset has a huge impact on the level of custom enjoyment of their cellular solution and thus their usage.
2) Customers however are cheap with handsets.
3) Customers are willing to pay more in monthly fees
So it makes sense for carriers to essentially lend customers money for the handsets as part of their monthly fees to get their total spend up. For every $1 in handset subsidies they get $3 in revenue which means $2 in revenue for the carrier related business.
That article was during 2012. During that year the subsidies on Android were $250-300 while the iPhone subsidy was $420. Of course Verizon wanted customers to buy an expensive dataplan and then kick back $8 / mo extra. But despite stores being mildly pro-Android 70% of their contract base still went iPhone when they could.
Note the May 2012 date. They went the other way. They increased subsidies. The article was wrong.
Let me clarify. Their addresses are going to be virtualized relative to the IPv6 space. The are going to look to the IPv6 customers as if they were on Verizon / AT&T... local routable subnet while in reality they are passing through a gateway. I'm not saying the servers are virtualized but the route will be.
Not they aren't ready. I've attended conferences on this. Verizon is getting close to ready. Comcast is done and it was 6 years of agony. They still have lots of stuff in their consumer / small business setups that makes v4 assumptions.
The rush is to be able to reclaim that valuable v4 address space to sell it at a premium to business.
No they aren't. There is a $40 / mo per smartphone fee and $17 of that goes towards subsidies. dumb phone users don't pay the smartphone subsidy fee.
10.6 wasn't a free upgrade. $20 is cheap it was during the time when Apple was lowering prices. They aren't defrauding anyone the prices on the upgrades are made public and haven't changed.
I have to say thought that is $20 bothers you, you aren't a good fit for Apple. You'll get hit far worse on other things. You will be much happier in the skinflint world of Linux or Windows.
OpenCL is a huge speed up and OpenGL 4.1 was too. If you go back Core Audio, Core Video, and Core Animation where huge speed ups. For a desktop (i.e. audio video...) I think there is a little doubt that OS X is way faster on good hardware than Linux.
We have pretty good data and the answer is absolutely yes it is. The overwhelming majority of OSX users never need to do a full OS reinstall. They move config from computer to computer. I'm not an average user but haven't done a clean install on my main system since OS 10.1.
If I hadn't posted already I would have moded you up. Very well said and completely true.
I'm going to agree with damn_registrars. Unless you are really young there is no such thing as a decent new bike for $200. The decent parts cost way more than that. You would be far better off going with a used bike and getting it cleaned up for that money.
If you were a dumb phone user you aren't paying the subsidy. The dumb plans don't include the smartphone subsidy cost.
That's not true. It was true in 2012 but not in 2013 or 2014. The carriers have been trying to prevent the formation of an Apple monopoly in the USA and pushed Android subsidies up to Apple levels. Apple is currently negotiating for yet another increase for the iPhone 6 so this may happen 4Q2014 but for about 18 mo it wasn't the case.
Once home / small business switch over the content providers are going to be virtualized. Which means that service will stop working, geolocation being the first to go. They'll lose the ability to meaningfully regulate traffic (everything is coming from West Virginia). It is fairly east to switch most websites over. Most consumer content will switch with a few years of the carriers being ready.
No they won't do that. Carriers have been pretty clear they aren't implementing carrier grade NAT and supporting it. ARIN has been hostile to them making use of carrier grade NAT. It isn't happening.
Oracle gave OpenOffice to Apache not Libre Office. Nothing changed about Libre Office. If you could recommend it before you can recommend it now.
I don't see any reason that's true. They are mostly the same. That's like saying Microsoft Office 2013 and Office 365 will make sure that neither is successful.
If you don't need document management then SharePoint won't do anything for you. It is entirely possible that a wiki would be a better choice. At this point sing you already have it maybe consolidate down to just a few SharePoint sites like: Engineering, HR, Accounting... That way you can use SharePoint for areas where people are unfamiliar with who to talk to. So for example if you don't know who in Accounting handles foreign checks and what the policy is for foreign purchase SharePoint works for you, but inside engineering you use something else.
Well Yammer is a new feature. People in large companies often don't know the other people in their teams. You may not have worked in large enough organizations where this is a problem.
Email is a 1-1 or small-small collaboration system. Document management is many-many.
Because that's not a SharePoint problem it is a SharePoint as you have it configured problem.
I don't know about dedicated but yes you need at least one person in each document creation group who sees SharePoint as one of their core job functions. The librarian is more at the level of integrating across the company. It sounds to me in your description like your company might be too small to have the sorts of problems that SharePoint is designed to fix.
No argument there. SharePoint is a serious lock in. The companies like yours who are using it wrong have much less of a problem than the companies who are using it right. It will cost those companies a fortune to get their data out. Most likely they will be on Share Point forever.
Yes that's fair. Though IMHO most companies and individuals don't really use any of their software right. So for example the #1 thing people do with Excel is keep lists. There is far better outlining software than Excel which is much better for lists. But...
I think most companies (over 100 employees) need a document library. Most companies write a large percentage of their documents in Office. I think Share Point is a reasonable choice for a document library for most companies. That's the general use. Then you throw in project resources like team calendars, team wikis... I think it is rather good and less hassle than most of them. But that's different than saying no hassle.
Taking the example of a file share. What most people do with a file share is better done in SharePoint than NFS. Which is not to say Share Point makes a good no frills file share but the changes that Share Point induces are the right kinds of changes for most companies and will benefit them.
Similarly on things like project calendar sharing.
I wouldn't say SharePoint sucks. It sucks when you don't use it as intended. The GP was using it as a file share. SharePoint is very good as a light document management solution with MS Office integration.
As far as the rest, I said the comparison is to Documentum. Documentum generally recommends an implementation team. It is not uncommon that Documentum has a permanent IT support department (i.e. multiple people). Share Point can get by with less than 1. As for metadata and using it its way. Yes. That's standard for a document management solution.
____
Imagine if someone was talking about SQLServer as a replacement for Oracle and we were discussing the lower cost and lower administrative hassle relative to Oracle. And you kept coming back with a comparison to using text files and how much easier and less of a problem text files were. SQL Server did a great job of taking Oracle administration (a skilled profession) and making it something that a moderately skilled admin could do part time on the side. Similarly SharePoint gets you 80% of Documentum for 20% of the cost.
1000 users on Documentum fully configured is probably just under $1m.
1000 uses on Sharepoint fully configured is about $200k
OTOH
1000 users on a fileserver fully configured is going to be $20k.
That's where you need to make the comparison. For smaller business, SharePoint online you can be at $7 / user / mo for a small business. There is nothing remotely like that for the higher end suites. But heck I can get 50g file share in the cloud for free.