Jump to content

Wikipedia:Reference desk/Archives/Computing/2014 January 17

From Wikipedia, the free encyclopedia
Computing desk
< January 16 << Dec | January | Feb >> January 18 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


January 17

[edit]

Wikia and templates and images, oh my!

[edit]

I would have asked this at the help desk, but I'm not sure if it's Wikipedia-related enough, even though I know Wikia uses MediaWiki. I'm attempting to add a category to a template (specifically, an image tag) so that all the images that transclude the template will put the image in that category. I found that adding it with includeonly tags seems to put the category name on the image's category list but does not actually seem to put all of the images in the category (a good 200 are listed but I know I tagged at least 500 by this point). Is there something I'm doing wrong or is it a lag of some sort? Thanks for any help. - Purplewowies (talk) 01:32, 17 January 2014 (UTC)[reply]

Yes probably a lag. The job queue is often pretty long you can see it at [1]. You could try doing a WP:NULL edit on some of the pages. This should recalculate the categories and you can use that to test the templates working correctly.--Salix alba (talk): 00:41, 19 January 2014 (UTC)[reply]

Net Neutrality

[edit]

Hello. I have been hearing a lot about net neutrality in the US. I was wondering, if the carriers are allowed to discriminate what traffic they allow, would it mean the end to hosting on a personal server? For example, if I have comcast and I go to a friend's house and she has some other carrier, will I still be able to ssh to my computer or will that not work anymore? Thanks! 24.128.61.100 (talk) 04:40, 17 January 2014 (UTC)[reply]

No answers yet, so I'll bite...
Today, many IP addresses are globally unique and globally routable. This is an emergent property: the internet protocol allows connectivity across multiple networks; and we have a situation in which many networks exist across the globe, and all choose to connect together.
Some day, it may be the case that many networks choose not to connect together. Or, they may provide connectivity at a very sluggish speed.
The Internet is a complicated piece of technology. As of 2014, nearly "everyone" agrees that there is one internet. Specifically, this means that we all defer to one organization (IANA) to provide canonically correct unique addresses. But the technology that underpins the internet protocol doesn't require that structure at all! I can make my own domain server and my own IP allocation server, and build my own network, and write my own routing tables, and if my network becomes large enough to become useful, I can choose to charge other people money for access to my network. That's how the commercial internet developed - a bunch of private networks started charging subscribers for access - and most users ultimately wanted access to government networks that the ISP's didn't even own! When the marketplace became more competitive, these networks needed more value-add; so as a bonus, they also provided connectivity to other private networks... and today, almost all the networks are connected. The competitors had to start cooperating.
At least a few economists and theorists believe this business arrangement is unsustainable. For example, I took a class with this guy, Ramesh Johari, who has extensively published his work on the topic of network graph connectivity in an internet that is dominated by private enterprises who are actually competing against each other. It is a fascinating problem in game theory, because the decision problem is quite complex. Any organization - an internet service provider - has more value if it is better connected to other networks. But if an organization has more value, it will demand more aggressive terms for its peering arrangements; and it will try to put its competitors out of business - resulting in fewer networks to compete with or to cooperate with!
So, it is actually possible that in the next decades, we will find that an Internet is not sustainable from an economic perspective. That is, it may be unprofitable to have one, global, unified network. It may be more profitable for each individual company to maintain its own large network that is not connected to many peer networks. To an end user like yourself, that means that some valid IP addresses would not be routable for you.
Now the hard part is, where exactly is "neutrality" in all this mess?
Network neutrality explains some of the many ways that the term is used. The biggest problem is, everybody likes the way "neutral" sounds, but few people can agree on which policy decisions are actually "neutral!" Does "neutral" mean that every network must play by the same rules? Or use the same quality of service? Does it mean that a government auditor must enforce packet-latencies according to some set of requirements? What packet latency requirements are "neutral"? If an organization buys new gear improving speeds within its own network, must it be required to buy new gear for all its competitors (i.e. everybody in the world) so that all network traffic flows equally ? These are farcical solutions to a very complex economic problem. It will take many decades for public and private sectors to figure out what to do.
So in the short run, nothing is changing, and that means your friend's commercial internet service provider probably will continue to be globally routable. I can assert with high confidence that the round-trip travel-time for a packet from your house to your friend's server is slower than the round-trip packet-time from your house to Google's server. Even if your houses are a few feet apart. Even if you live nowhere near a Google data center. Carriers, phone companies, internet providers, and your home's Wifi router all already discriminate in favor of certain types of traffic. So we're living in a "not very neutral" internet already, but there has not yet been a loss of graph connectivity.
Nimur (talk) 15:35, 17 January 2014 (UTC)[reply]
Oh, here's a new book, available at no cost: Economic Modeling in Networking, which sets out to answer questions like "what's the difference between efficiency and fairness," in the context of internet economics. Nimur (talk) 15:40, 17 January 2014 (UTC)[reply]
Thank you! That was very helpful! 24.128.61.100 (talk) 02:54, 19 January 2014 (UTC)[reply]
Consider what happens to a site like Wikipedia. Without net neutrality, either Wikipedia pays money to the ISP's - or their slice of the bandwidth is reduced (presumably heavily - or else nobody pays the premium). Since Wikipedia doesn't have a whole lot of money to bid on the available bandwidth compared to (say) YouTube - they'll inevitably wind up running really slowly by comparison. Worse still, there isn't (yet) one single place you'd have to go to in order to negotiate a payment. In principle, you'd have to negotiate separately with every cellphone provider, every cable TV network, every telephone-DSL provider and so forth. To reach global coverage, you'd need a massive army of negotiators working with ISP's around the world. You'd also have to figure out where it's worth putting the money. Do you fund good coverage to AT&T and let Sprint customers get slower coverage? Do you provide better coverage to West-coast US users or East-coast?
Worse still, if one service provider decides that they can make a lot of money by charging web site owners money - it's only a matter of time until someone realizes that they can offer consumers "free" Internet and live off of the fees that web site owners pay them. We know that "free" or "very low cost" services are popular to consumers - that's why our (free) broadcast TV shows are stuffed full of annoying adverts. So once one ISP goes that route - it's only a matter of time before the others do too.
In a world where most people use "free" internet and web sites pay the bills - how can independent web sites possibly survive? Wikipedia couldn't function in that world - so it would pretty much die without taking on advertising, paid promotional articles or switch to a pay-per-use model.
Pretty soon, the Internet starts to look like TV. Advertisers will start to drive content. TV shows are not made to make viewers happy - they are made to pull in advertising revenue. That's how the Internet will eventually end up. Gone will be the time when a bunch of enthusiasts can put something together in a garage and become the next Google - they simply won't be able to break into the closed-society of high-paying web sites - so the general public will simply never see them.
None of those consequences are very nice...each one destroys the beautiful thing that the Internet has become.
SteveBaker (talk) 17:22, 20 January 2014 (UTC)[reply]
Thank you Steve! — Preceding unsigned comment added by 24.128.61.100 (talk) 00:06, 21 January 2014 (UTC)[reply]

The IPv4 endpoint of a 6to4 IPv6 address

[edit]

Hello, I have a dynamic IP that I have been using for years. When the internets went all magic and changed to IPv6, my IP starting messing with me. I never received a IPv6 address for about a year, then started getting them sporadically. Recently, I seem to get one about once every two weeks (I receive a new IP every 8 hours when my dial-up IPS disconnects me, I leave my connection always on and have it auto-re-dial when disconnected). About once a month, I receive an IP that flips back and forth from IPv4 to IPv6; I know it is the same IP because the WHOIS of a IPv6 lists my IPv4 address and I know it flips back and forth because I use "special:mytalk" as a sandbox (so I could be on my IPv4 talk page/sandbox and then an hour later save an edit and it goes to my IPv6 address which WHOIS traces back to the same IPv4). So my questions are: 1) why do I receive an IPv6 address only some of the time, and 2) why then does it flip from an IPv4 address only some of that time? Rgrds. --2002:4055:D812:0:0:0:4055:D812 (talk) 17:09, 17 January 2014 (UTC)[reply]

You must have an IPv4 address all the time otherwise a big chunk of the internet wouldn't work for you and you couldn't use 6to4. Also it sounds like you are conflating issues which makes it difficult to diagnose anything. You could be connecting to wikimedia servers by IPv4 because your connection currently has not managed to obtain an IPv4 address by any means it's set up to or you could be connecting to wikimedia servers by IPv4 because you browser for some reason chose it over IPv6. Rather than confusing yourself with how you connect to wikimedia servers, you should first better diagnose whether you have a working IPv6 connection all the time or what.
For example, try pinging a host on IPv6 (ping -6 ipv6.google.com). Even better, work out from you router or whatever else is getting the IPv6 address whether it always has a public IPv6 address or what. Once you've worked out the state of your internet connections IPv6, you can then either move on to trying to work out why you sometimes connect to wikimedia servers via IPv4, if it really matters; or work out why your internet connection doesn't always have IPv6 (a problem with the relay or your router perhaps). At a minimum, you should at least test with another host, e.g. try accessing [2] and [3].
BTW, the internet never went all magic and 'changed to IPv6'. Sadly, many servers let alone connections still don't have IPv6. But some have had IPv6 for a long time. I started using IPv6 via a tunneling service in ~2007 which was in some ways fairly late (although root servers only got IPv6 in 2008). You may be thinking of World IPv6 Launch Day in 2012 when a number of major websites (including content mirroring providers/CDNs) added AAAA records for their main websites with the intention of leaving it on permanently. This meant these services were accessible by IPv6, so those with IPv6 may have accessed them via IPv6 depending on the set up of their browser and/or OS.
But many services did not take part and as I've said, a number already had AAAA records by that time so were alreadt accessible via IPv6. And a large number of connections still don't have working IPv6 so are still using IPv4 hopefully without issue. (One of the reason it took so long for these major services to enable IPv6 was not so much because their backends couldn't handle it, but because they were concerned that misconfigured hosts with IPv6 that wasn't working properly would have problems accessing their services. Google for example had ipv6.google.com for a while. And they even had a way of adding AAAA records for their services even before the IPv6 (trial) day, see IPv6 brokenness and DNS whitelisting.)
Nil Einne (talk) 13:26, 18 January 2014 (UTC)[reply]

Changing default zoom in Pages 5.0

[edit]

I'm using Pages 5.0 in Mavericks, and I want to change the default zoom to 150%; does anybody know what I need to type into Terminal to do this? I've seen people suggest the command "com.apple.iWork.Pages SLDefaultsPageScale 1.50", but of course that only works in Snow Leopard. --Lazar Taxon (talk) 19:17, 17 January 2014 (UTC)[reply]