Wikipedia:Reference desk/Archives/Computing/2014 April 10

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.

April 10

Will a standard internal optical drive fit in this bay?

Will an internal optical drive that is designed to be accessed externally (like most of them) go in a bay like this one in a Dell XPS 8300, with a door over the drive and a button on the side? Photo shows drive door open and Photo shows drive out with my finger on the open-close button. Bubba73 ^{You talkin' to me?} 04:30, 10 April 2014 (UTC)

It looks pretty standard to me, but it would be more useful to have a pic with the door closed. The drive should be 5.25" wide and 1.6" tall if I remember right. Based on web search finding the xps 8300 is a recent machine, it's likely to take a standard Serial ATA optical drive (older computers used Parallel ATA). By comparing the pictures of the connectors in those articles with the one on your drive, you should be able to figure out which type you need for sure. 70.36.142.114 (talk) 05:13, 10 April 2014 (UTC)

The motherboard has SATA connections. It looks to me like I will have to take off the door from the case to get to the button on the new drive, unless the button on the side somehow will work on the new drive. Photo with the door closed Bubba73 ^{You talkin' to me?} 05:22, 10 April 2014 (UTC)

Well, if the button on the case will send a signal to the new drive to open/close, then it will work. The drive that needs to be replaced doesn't have a button on it. So the issue is whether or not the case switch will send the new drive the open/close signal. I suppose it must. Bubba73 ^{You talkin' to me?} 05:43, 10 April 2014 (UTC)

Oh I didn't understand what you meant about a button on the side. You mean some kind of remote button wired from the case? I think you should open up the box and see where the wire goes. Am I correctly understanding you're trying to replace this drive? 70.36.142.114 (talk) 06:41, 10 April 2014 (UTC)

Yes, I am replacing the drive that is in there - it won't write past a certain point. The button on the case is not a mechanical button that presses the button on the drive - it must be an electrical button that sends an open/close signal. I'll look inside. Bubba73 ^{You talkin' to me?} 15:26, 10 April 2014 (UTC)

I should clarify the way it works. Pushing the button on the case does not open the door - it tells the drive to open and that pushes the door open. When you press the button again, the drive goes in and a spring pulls the door closed. Bubba73 ^{You talkin' to me?} 17:03, 10 April 2014 (UTC)

Well, it turns out that the one in there was a standard one - I just couldn't see the button from the outside. I replaced it with a new drive and it fixed my problem writing to disks. Bubba73 ^{You talkin' to me?} 20:39, 12 April 2014 (UTC)

Resolved

Phoning Australia

This question is not asked from the perspective of someone such as JackofOz. I'm an American :-)

Friends of mine moved from the USA to Australia some years ago, but they somehow have a setup whereby I call their old Indianapolis phone number and it rings in Frankston. How does this kind of thing work? They recently sent a bunch of American acquaintances an email (basically "remember the time difference!") and mentioned VOIP, but I can't understand how the two are related (this is a real phone call, area code 317, not a Skype call), let alone how I can dial an Indiana phone number and get someone in Victoria. Nyttend (talk) 06:18, 10 April 2014 (UTC)

VOIP operators such as Vonage offer a variety of services whereby a 'real' telephone number is ported onto their system (much like if you switch landline or mobile provider and wish to retain your number). Inbound callers then ring a 'local' number which is routed through to the service provider who transmit the call on their own network (in this case, over the internet). The call can be received on an internet enabled device anywhere in the world. Some providers even offer interface units to allow the reciever to use a normal telephone. Likewise Virtual numbers are available, where you can choose a number in any country for people to ring you from (as used heavily in call centres and support lines). So if you wish you could have a 'local' number in 5 countries that redirect to a VOIP phone in Australia, with the callers only paying local rates. Nanonic (talk) 07:00, 10 April 2014 (UTC)

I'm guessing the OP doesn't really understand what VoIP is anyway. If your phone last mile is fibre, it very often will be partially VoIP nowadays even if it's not marketed as such and is designed to be transparent to the end user. I think even a large percentage of fibre backbone including that used for voice is entirely IP, although I'm less sure of that. (Similarly LTE (telecommunication) is supposed to result in a final move to an all IP packet switched network although the actual implementation of voice over LTE seems to have been slow.)

Our article on VoIP isn't brillant but it does seem to cover the basics including the fact that VoIP is voice over internet protocol, not voice over internet. What people would call the 'internet' doesn't have to be involved at all in VoIP. If I call someone on a LAN, it could be the internet isn't involved at all, in fact there could be no internet connection at all but I would still likely be using VoIP, unless for some reason I'm really not using IP which is unlikely for any simple LAN and probably most complex ones.

Note also the Skype comment seems confused. Even Skype is able to provide real phone numbers in a number of countries, see e.g. [1] [2] [3]. (I don't know if Skype supports porting in any country but the OP seemed to think you couldn't have a real phone number with Skype.) And of course people have also been able to Skypeout to a real (Public switched telephone network derived) phone number in many countries for a while.

Nil Einne (talk) 13:25, 10 April 2014 (UTC)

You didn't mention if this is a landline or a cell phone. Cell phones can operate over various area codes. One of my family members lives in Mexico but still uses her California area code. — Preceding unsigned comment added by Justin15w (talk • contribs) 14:40, 10 April 2014 (UTC)

Call forwarding. 70.36.142.114 (talk) 19:58, 10 April 2014 (UTC)

Response to two separate people: (1) The Indiana number is a landline. I think it's a landline in Victoria, since I've called them several times, and it always seems to be answered by a different member of the family. (2) Nil Einne, I really don't understand VOIP much at all. Before asking this question, I decided to read the VOIP article, and all that happened was that I got more confused. Among other things, I didn't notice that it didn't require Internet usage. I definitely didn't realise that you could have a real phone number with Skype; I thought it was something you downloaded from their website (with payment if you want extra features) and installed on your computer, and nothing more. Nyttend (talk) 01:03, 11 April 2014 (UTC)

Architectural style

What does architectural style in informatics mean? Is it a synonym to architectural pattern, so that simply a hatnote in architectural style should be added? Or is it more complicated? --KnightMove (talk) 08:05, 10 April 2014 (UTC)

Windows XP - gone slow, no patch forthcoming

I have a Dell OptiPlex GX520 (the BTX Small form Factor (Matrix Cypher Case) - it looks very much like this) with Pentium 4 running at 3 GHz with 3.6 GB hard disk and (allegedly) 1 GB RAM, Windows XP Pro SP3.

Microsoft announced some time ago that they would stop supporting Windows XP. They did this following the patch update of 8 April 2014. After that update (which contained five separate patches), I rebooted as advised, since when my PC has run slower; sometimes the keyboard or mouse seems to stop responding, only for whatever I've typed/clicks I've made to suddenly appear a few seconds later. Some applications are noticeably slow to launch. Has anybody else reported this? Given that Microsoft won't be fixing it (even if it is a bug in the latest patches), what's the way forward? --Redrose64 (talk) 13:16, 10 April 2014 (UTC)

Linux. All the best, Rich Farmbrough, 17:47, 10 April 2014 (UTC).

Your slowdown may be from a virus infection. Don't try to cure it. Ditch XP. Use Windows 7 if you have to stay with Windows, Linux otherwise. Stay away from Windows 8. 70.36.142.114 (talk) 19:56, 10 April 2014 (UTC)

Windows 7 won't fit. It takes 8 gigabytes on the HDD, bare minimum. - ¡Ouch! (^{hurt me} / _{more pain}) 07:29, 11 April 2014 (UTC)

My mistake. The HDD is 36 GB, not 3.6 GB - long rows of unseparated similar digits. --Redrose64 (talk) 13:13, 11 April 2014 (UTC)

Windows 8 is actually known to generally perform better than Windows 7 with borderline configs (actually most comfigs but it's less noticeable). That said, the cost and complexity of moving to either is probably not worth it on such an old PC unless you happen to have a free or very low cost option available (e.g. uni MSDN access). Some *nix may be worth considering. Nil Einne (talk) 18:05, 11 April 2014 (UTC)

If you haven't already done so, first try a reboot.

If your PC started running slower immediately following those patches, then they would indeed appear to be the culprits. If you have a checkpoint you can go back to, that might be worth trying (save a checkpoint as is now, in case you ever need to go back). Also, my understanding is that "no longer supporting XP" just means no more future patches will be released, you can still go to their web site to get the old patches. So, you might be able to go to a previous version that still worked well.

Increasing the RAM might also fix the problem. Download and run the http://crucial.com Crucial Memory Advisor™ Tool to see what you can expand it to, and about how much it will cost. StuRat (talk) 13:21, 11 April 2014 (UTC)

I rebooted about 3 times since. It now seems to be operating at approx the previous speed, but began doing so after the second reboot. The third reboot was because the whole machine suddenly stopped talking to my broadband router: Firefox, Chrome etc. weren't retrieving pages - they showed "Waiting for [domain]" with the spinny thing going on far too long. I tried pointing my browser at http://192.168.1.254/ (that being my broadband router's IP address on my LAN) and even that wouldn't display anything. At command prompt, wget http://192.168.1.254/ brought nothing back either, so I rebooted after which it has behaved. Curiously, the router responded normally to ping 192.168.1.254

Theory: Microsoft want people to lay out cash on the latest Windows, so they've introduced deliberate WONTFIX bugs into XP to force users to stop using it. --Redrose64 (talk) 15:01, 14 April 2014 (UTC)

Is the heartbleed bug something we should be paranoid with?

Is the whole commotion around this bug just a media hype? As I understand it, an attacker can only get 64k, fetched at random, and it's improbable that someone discovered it before the researchers who reported it (and apparently acted in good faith). There are certainly other non-addressed security flaws out there: ebay does not use https, Windows XP is not being maintained anymore, people still click on links received by email from an unknown source, and all sort of weak passwords. OsmanRF34 (talk) 13:19, 10 April 2014 (UTC)

There has been discussion at WP:VPT#Heartbleed bug?. --Redrose64 (talk) 13:21, 10 April 2014 (UTC)

The discussion there primarily seems to concern the possible effect of the bug on wikimedia sites, not so much the more general effects. I doubt a wide spread discussion of such will be particularly welcome there either.

Anyway as for the OP, I don't think it's completely improbable someone with malicious intents discovered it before the researchers did. More to the point, now that it's out there, anyone who wants to use it for malicious purposes has probably had little sleep gathering as much as they can from the remaining unpatched servers. 64k may not sound like a lot, but as I understand it it's actually often not that difficult to hammer a server repeatedly without much risk of being blocked. And of course anyone who seriously wants to take advantage of this likely has large zombie nets they can use anyway.

While there are obviously plenty of other security risks out there, this is particularly widespread and basically can have a significant effect on the end user without the end user having much control, or at least not in a simple or obvious manner.

If you're correctly concerned about the security risks from Windows XP, the vast majority of people have the option of just not using it, at least on an internet connected PC.

If you're worried about eBay not using HTTPS, either don't use eBay or don't reuse your eBay password somewhere else particularly not on PayPal or some payment gateway (worst case scenario someone compromises your account and destroys your profile by selling stuff and not delivering or bidding on stuff but not paying).

Okay you can avoid any servers with the bug, but you have to test them first and it could easily be a bit late now. There have been somewhat similar problems in the continual stream of servers leaking their user accounts and passwords and occassionaly credit card info due to poor security but these are generally only one or two major sites at a time. (Edit: Credit card info can also come from other sources such as poorly secured payment machines or processing networks.)

And getting back to my earlier point, when did you first hear of this bug? Some people suggest people stay away from the internet, or at least anything where they want some degree of confidence people aren't going to snoop on their activity given the difficult knowing precisely what's at risk and what isn't. I think this may be a bit extreme but in any case, by the time someone heard this it may have already been too late. E.g. Yahoo was at risk, IIRC by the time I heard about the bug they had already patched as had wikimedia, but this probably wasn't the case for those with malicious purposes.

Incidentally, provided someone isn't using Tor or other networks with an extremely untrustworthy middle, probably security agencies are the biggest risk for man in the middle style attacks in general. So yes while eBay's alleged practices may not be ideal and there may be a few who either get lucky or have people in the right places, it probably isn't as risky as it seems. Yet a lot of the risks of the bug don't require anyone in the 'middle', my connection to the server can be completely trustworthy but the server may still give away stuff I don't want them to other people with malicious purposes.

OTOH when it comes to MITM, there are also some other risks (of unclear frequency) that are somewhat different from the norm, e.g. a leaked private key and the risk of that to both current traffic but also any past traffic that's been recorded if Perfect Forward Secrecy isn't used. (Obviously there is always some risk that the private key is going to be exposed at some stage, but it may not have been thought so many as may or may not happen here depending on how likely such a scenario actually is with this bug.)

In other words, whether this bug is really as bad as some, including some security researchers are making it out to be, I can't say for sure but it definitely doesn't seem like a run of the mill thing.

P.S. I forgot to mention but another thing about this bug, as I understand it, it's difficult for anyone to know if it's been exploited unless people turn up with stuff which they can only presume was found from an exploit. You may be able to test and work out what could have been gathered from successful exploitation and I guess you could also implement something to try and monitor current exploit attempts now that it's known, but that's about it. This compares to many other exploits where often there's something in some log, even if the admins only realise it too late.

P.P.S. Probably the biggest thing against this bug being useful is not really the size but the 'random' but. And not that the info that may be revealed may not be useful, but that it may not be easy to automatically parse all you find to actually get the useful info. Stuff like credit card numbers and username/passwords could potentially be picked automatically. Trying to find and assemble a servers private key, even if it is revealed may not be so easy. It's the sort of thing that some manually work may be worth it if you're interested in that sort of thing, but as with most things, it's a matter of scale.

Nil Einne (talk) 14:08, 10 April 2014 (UTC)

I think that extracting a private key from the leaked data would be easy, even if there is no identifying prefix or suffix. You have the public key, and testing whether a private key matches a public key can be automated, so you just test every substring of the appropriate size. -- BenRG (talk) 22:14, 10 April 2014 (UTC)

As with the Target breach, I think there's a low probability that it will harm you personally, but it's worrying. In terms of potential impact, I think it's the worst security failure in the history of the public Internet. It is easily exploitable, it has affected an enormous number of servers for up to two years, and there is probably no way to tell whether it was exploited in that time. Bruce Schneier ranked it 11 on a scale of 1 to 10. The actual impact is hard to guess. It could be limited to the current IT staff all-nighters. The biggest problem is the ongoing loss of trust. You can't be sure that an affected server didn't have its private key stolen. Anyone who stole the private key can impersonate the server (MITM), even if the server no longer uses that key. To prevent this, the old keys or certificates have to be revoked. But the number of affected servers is enormous: the revocation list would be too large to ship with every SSL/TLS library. We might end up with everyone having to use OCSP. And there's no way for a customer of a web site to be sure that the site's operators have revoked the suspect keys as they're supposed to. It's a terrible mess.

I suppose someone could write a browser extension that popped up a warning if a site has a certificate issued before April 8, 2014. Except that according to this post, at least one CA is issuing replacement certificates with the same valid-from date as the compromised certificates. Great. -- BenRG (talk) 22:14, 10 April 2014 (UTC)

Ebay's login page definitely uses https. Yes, Wikipedia's own replacement certificate (per VPT) has the same "not before" as the old one. It's possible that certificate pinning or the EFF SSL observatory might pick up on changed certificates. Regarding MITM, it is not needed, most sites (including Wikipedia) still use SSL cipher suites without forward secrecy. So if you can get the private key and also monitor the site's web traffic, you can decrypt all communications passively, from before and after you got the key. That said, the entities most capable of doing this are three-letter agencies who probably aren't going to try to impersonate you on Wikipedia even if they know your password. 70.36.142.114 (talk) 02:23, 15 April 2014 (UTC)

AutoCAD

I am using AutoCAD LT 2012.

I have a model in model space and extra stuff in layouts in paper space.

Once I have got the viewports setup, is there a way of freezing/locking the model so that I can't edit it through the layout viewport, even when on the MODEL sub-tab. To edit the model I want to have to leave the layout page and go back to the MODEL page.

I feel convinced I've come across a drawing setup to do this, but I don't know how.

I know it could be done by using separate layers on models and layouts and locking the model ones but this feels a bit cludgy.

Thx. -- SGBailey (talk) 14:39, 10 April 2014 (UTC)

You want to lock the viewport, to do this - select the viewport itself, open its properties and set Display Lock to Yes. In some versions of AutoCAD you can do this in paper space by clicking the padlock at the bottom left of the screen. Nanonic (talk) 06:27, 11 April 2014 (UTC)

Python Global Variables

I'm in a spot of trouble with global variables. What I want to do is something like this (this is just a skeleton of my entire code):

class Stream_Tweets():
	tweet_database=[]
	tweets=[]
	def Streaming(duration):
		#function that streams tweets for duration,details redacted
		#global variable stores streamed tweets in a list
		global tweet_database
		tweet_database.append(tweet)
	def Cleaning_Up():
		#function that cleans up tweets,repetitions, removes hashtags, etc
		#global variable gets updated to store only cleaned up tweets
		global tweet_database
		tweet_database[i]=clean_tweet
	def Sentiment_Analysis():
		global tweets
		for tweet in tweet_database:
			tweets = #store the tweet, and the sentiment score

After I wrote this, there were errors saying something about unbound functions for Cleanin_Up() and Sentiment_Analysis(), so I added a @staticmethod decorator at the top of each. But now I'm getting errors that say "global name 'tweets' is not defined" or "global name 'tweet_database' is not defined. The code appears to work, however, if I define the global variables outside the class Stream_Tweet(). But in that case, when I run the following snippet multiple times:

f=Stream_Tweets #Object 
f.Streaming(100) #Stream for 100 seconds
f.Cleaning_Up()  
f.Sentiment_Analysis(10)

Instead of the database getting rewritten everytime I stream fresh tweets, the new tweets get appended to the old ones. So, to sum up, can someone explain, in very layman terms what staticmethods are (I've googled it, but I still don't get it), and how to make it so that my global variables are declared inside the class, are available for all the functions inside the class, and everytime a new object is created, the values of the global variables are initialised to [] ? Much thanks for your patience. La Alquimista 14:55, 10 April 2014 (UTC)

You don't need global for any of this. "static" stuff is shared between all instances of the class; nonstatic (aka "member") stuff has its own copy for each instance, which is referenced through the "self") parameter:

#!/usr/bin/python

class Stream_Tweets():
        tweet_database=[] # static variable - shared between instances

        def instance_method(self, tweet):
                # instance methods can access static content too; it's still static, even though its accessed through self
                self.tweet_database.append(tweet)

        @staticmethod
        def class_method():
                # this is a static method, so one uses the class name to access static storage
                print Stream_Tweets.tweet_database


# create two instances of Stream_Tweets(), and show that they share the same internal tweet_database,
# because it is static
a = Stream_Tweets()
a.instance_method("uno")

b = Stream_Tweets()
b.instance_method("dos")

# print the results
b.class_method()

-- Finlay McWalterᚠTalk 15:15, 10 April 2014 (UTC)

Maybe this will clarify things for you (and maybe it'll make things worse...): add

print id(a),id(b),id(a.tweet_database),id(b.tweet_database)

and you'll get something like

     10568768 10568840 10438920 10438920

note that a and b have different ids (they're different objects) but a.tweet_database and b.tweet_database have the same id - it's the same object with two different means of accessing it. -- Finlay McWalterᚠTalk 15:24, 10 April 2014 (UTC)

I'm beginning to see it now. I don't get the "self" thing though. Python has dynamic typing, so I guess when it sees self.tweet_database.append(), it understands that self is an object and thus, we can call a function like Example(self) by using a.Example(), that is, without adding parameters inside the (). However, suppose that there's a function which has self in its arguments list, but doesn't use it as an object anywhere inside the namespace. In that case, when we call that function with specifying any parameters within brackets, will Python throw up an error like "this function takes 1 argument"? La Alquimista 16:38, 10 April 2014 (UTC)

Yes, it will. There's a special rule (one might call it a hack) in Python that when an attribute is looked up, and it's found in the object's class (not in the object itself), and it's callable, and not a static or class method, you get back a wrapped version of the function that inserts the object as the first argument when you call it. E.g., foo.bar returns the equivalent of partial(class(foo).bar, foo), with the result that foo.bar(x) works out to class(foo).bar(foo, x). The function-call parentheses don't need to be there: you can assign foo.bar to a variable and call it elsewhere like any other function object and still get the implicit self argument, which is often useful. -- BenRG (talk) 18:10, 10 April 2014 (UTC)

(I think I misunderstood your question when I wrote "yes, it will". The magic I described above is the only magic. Python doesn't look at the number of parameters or their names. Calling the first argument "self" is just a coding convention.) -- BenRG (talk) 18:57, 10 April 2014 (UTC)

You mean type(foo), not class(foo). --Tardis (talk) 02:58, 15 April 2014 (UTC)

Make sure you understand the self convention before proceeding further. BenRG explained but if there's still any unclarity, go ahead and ask further questions about it rather than remaining confused. As a subjective stylistic matter, I'd advise avoiding either global or class-level static variables if you can. If you want instances to share a database, pass the database into the instance constructor every time you make a new instance. That makes the program more flexible and easier to test, since you can instantiate multiple databases (including mock objects) if the need arises, and there's generally less hidden magic around. 70.36.142.114 (talk) 20:05, 10 April 2014 (UTC)

Another thing worth knowing about is the following unfortunate behavior:

>>> x = 'global'
>>> class Foo:
...     x = 'class'
...     @staticmethod
...     def bar(y=x):
...         return x, y
...
>>> Foo.bar()
('global', 'class')

Expressions in the direct scope of the class statement (such as the default expression for y) can refer to the class's x as x. But in subscopes, the class scope is simply ignored, and you will see a binding from an outer function or global scope if there is one. Python is designed in such a way that scoping can be resolved at compile time, which means that implementations could reject such code with an error like "references to Stream_Tweets.tweet_database must be qualified", which would have been helpful in diagnosing your original mistake and even more helpful in preventing the subtle bugs that can arise when both class and global bindings exist. But there are probably programs out there that rely on the current behavior, so it can't be fixed. Until Python 4. -- BenRG (talk) 00:02, 11 April 2014 (UTC)

outlook 2010 spellcheck two different languages in same message

i cant get outlook 2010 to correctly spellcheck when writing a message in two different languages. by default it spellchecks in english, and i write in english and french in the same message, so the french is spellchecked as english (if that makes sense). e.g. this is message: "Bonjoure mes garcins. Helli thare." Spellcheck corrects "Bonjoure" and "garcins" as thinking they are English words, when really they are french, but it correctly corrects "helli" to Hello" and "thare" as "there". And yes i do have "detect language automatically" checked.Billybob007 (talk) 15:45, 10 April 2014 (UTC)

.xls file password lost

A protected .xls file is telling me the password I enter is wrong. The password I enter is what I remember and have used for a while, and it is also recorded in a .doc file on a thumb drive where I keep copies of several important files, including this one. The copy on the flash drive (updated March 31) also says I have the wrong password. Is there a way to bypass the normal password protection and get into the protected file? --Halcatalyst (talk) 17:27, 10 April 2014 (UTC)

Never mind -- now it works. --Halcatalyst (talk) 17:41, 10 April 2014 (UTC)

Keeping the password on the thumb drive where you keep the document is ... unusual security practice. All the best, Rich Farmbrough, 17:54, 10 April 2014 (UTC).

Odd issue after updating to Ubuntu 13.10 from 12.10-Keyboard/Mouse not working

After updating to 13.10 from 12.10 there is a problem with the OS. It bots in to the Unity desktop, but with no cursor and no mouse. I can't even use the keyboard to open a terminal. It also says in a bubble to the side that my wireless has been disconnected.

I believe the problem here is that some time before the update I installed a Nvidia proprietary driver as I am using a GTX 660.

If I use the USB recovery to remove the driver, what packages do I need to purge/reinstall? Or is it not the driver? KonveyorBelt 17:33, 10 April 2014 (UTC)

JSON

In JSON, tweets look something like this:

{"created_at":"Thu Apr 10 17:24:21 +0000 2014"
"id":454308917812531200,"id_str":"454308917812531200"
"text":"RT @Bihar_ke_lala: blah blah blah"
"source":"web"
"truncated":false,
"in_reply_to_status_id":null,
"in_reply_to_status_id_str":null,
"in_reply_to_user_id":null,"in_reply_to_user_id_str":null,
}

If a tweet is actually a retweet of someone else's tweet, then the tweet might be truncated in the main "text" field, (even though the "truncated" field will still show a "false" value). Instead, the main text of the original tweet can be found in a new field called retweeted_status. This field appears only in retweets, and is absent in other tweets. I'm trying to check for this field in Python, and have tried giving something like this:

if data[retweeted_status] != None:
       #get the text of the original tweet

However, this doesn't work and the code still throws up an error whenever a tweet with no retweet_status field is encountered. What might be a possible solution for this? La Alquimista 18:31, 10 April 2014 (UTC)

You probably want something like

    if 'retweeted_status' in data:
        text_of_the_original_tweet = data['retweeted_status']
        ...

or

    try:
        text_of_the_original_tweet = data['retweeted_status']
    except KeyError:
        ...

or

    text_of_the_original_tweet = data.get('retweeted_status')
    if text_of_the_original_tweet is not None:
        ...

Note the quotation marks around 'retweeted_status' (I assume that was just a typo on your part) and is not None instead of != None (either one works, but "is not" is faster and more idiomatic). -- BenRG (talk) 19:17, 10 April 2014 (UTC)

What's up with AT&T Uverse internet?

I've bounced around between a bunch of cable companies for Internet access over the years...this time around, it's AT&T's turn. I've had very, very bad experiences with AT&T before - so I was nervous about signing up with them again...and it seems that my worst fears are well-founded.

Sadly, what I usually do (plug my WiFi router into a wired ethernet port on the cable modem provided - and get to work) isn't working. Firstly, their system comes with it's own WiFi router - I was all for unplugging it and using my own (just like I always do) - but the AT&T tech says that this won't work - and couldn't/wouldn't elaborate on why not. The official AT&T WiFi router seems to work OK on my wife's laptop and with android phones/tablets and with our Roku contraptions. But my Ubuntu (v12) desktop computer isn't happy.

With a wired connection to my desktop machine, I get connectivity to the router and I can fiddle with it's internal settings - but when I try to get out to the Internet, nothing. So I bought a USB WiFi dongle - and with that alone, my Ubuntu machine claims to have network access - but every site I try to reach times out. If I connect BOTH Wifi *and* wired networking - I do get patchy access to the outside world. The system will load a page here and there - but very often just sit waiting for minutes at a time.

Interestingly (and as a matter of desperation) I went in and changed some of the settings on the AT&T box. One thing was that I was sick of typing in the 10 random-character password they gave me - so I changed it to something more memorable. This worked fine until I had occasion to reboot the thing - and to my horror, it reverted to the original (factory set) password again! Several other settings that I've changed are also reverted when the machine is powered off and on again.

Surfing around for answers (especially googling "Ubuntu Uverse") gets a lot of other people with similar problems - and not one single good solution to it.

AT&T's service guys are incredibly unhelpful - they say that Linux isn't supported and that's that. (I protest that my Android phone & tablet work fine and those are supported - but the tech is too stupid to know that Android is Linux! Ditto with the Roku.)

So what's up with this thing?

SteveBaker (talk) 19:23, 10 April 2014 (UTC)

Interesting. I've had plenty of problems with AT&T, but I haven't had any problem running a desktop with Ubuntu 12.04 with a wired connection to an AT&T 2Wire router. I didn't need to do any fiddling at all with internals. (What has been my experience is that their phone support people are clueless. They clearly work from a script and will often give advice that is obviously wrong.) Looie496 (talk) 02:40, 11 April 2014 (UTC)

Agreed on their phone service people being useless and working off a script. I called them when my cable was out, read them the error message coming from the cable box and displayed on my screen, then they started by asking if the cable box was plugged in. "No, it's magically displaying error messages on my screen without being plugged in !" StuRat (talk) 03:40, 11 April 2014 (UTC)

Yeah, when I had AT&T Mobility, I was unable to edit a particular wiki because a vandal in the same IP range ticked off the admins enough that they put in a range block, and when I them to complain that someone else's terms of service violations were preventing me from doing what I want to do on the internet, the call center person didn't even know what a terns of service violation was. They went and got tier two CSR, and the tier two tryed to say that the IP range I was in with their service wasn't their's. What a bunch of boloney. People can say what they want about CenturyLink's customer service, but I'll take there's over some of these other companies' customer service any day. 71.3.50.250 (talk) 20:02, 12 April 2014 (UTC)

mAh with regard to portable backup battery

This is probably a dumb question but I just want to be clear about something. I'm going on a few vacations this year so I need a portable battery backup for my phone. I'm assuming the ones with more mAh are preferable, but I'm wondering if there's any way to determine how many charges a full one can last. My phone, a Samsung Galaxy S III, has a 2100 mAh battery. Does that mean a portable battery with 2100 mAh could only charge it once? For instance I'm looking at one with 10400 mAh - does that mean it could charge it roughly 5 times before it needs to be charged as well? Thanks. NIRVANA2764 (talk) 20:29, 10 April 2014 (UTC)

I believe that's roughly right, in reality, you probably won't get that much out of it, especially if it's a cheap battery, since they probably try to inflate the figures. Vespine (talk) 23:13, 10 April 2014 (UTC)

Replacement batteries for a cell-phone are not very expensive. You could consider buying 5x, if you need 5x a full battery before you can reload them. OsmanRF34 (talk) 00:13, 11 April 2014 (UTC)

Probably more like 3 charges from the 10400 mah battery since the charging and conversion circuits aren't 100% efficient. Look at newegg.com daily deals for a few days, they have those things on sale pretty regularly. 70.36.142.114 (talk) 01:44, 11 April 2014 (UTC)

There's some loss during the recharging process -- 10% to 25% loss is typical. I'd estimate that a 10400 mAh battery could recharge a 2100 mAh battery four times, not five, before running out of power. --Carnildo (talk) 01:44, 11 April 2014 (UTC)

Where are you going on vacation that you can't recharge ? If you can't use a regular wall charger because the wall outlets will be different there, how about a car charger ? (Are car cigarette lighters the same around the world ?) A car charger might be cheaper, more portable, and last longer than extra batteries. StuRat (talk) 03:46, 11 April 2014 (UTC)

Regardless those batteries can be quite handy, on the road, on a train trip, on the plane for 24 hours, trekking in the countryside, even if you have a hotel you might not be there long enough to fully charge your phone before you head out for adventures again. Vespine (talk) 04:40, 11 April 2014 (UTC)

Note there is a 4600 mah Mugen Power battery (mugen.co) available for the Galaxy S3... it's about 2x thicker than the usual one and comes with a replacement battery door to accomodate the extra size. 70.36.142.114 (talk) 10:42, 11 April 2014 (UTC)

Also note that while there are many many low-cost no-name models available from China on ebay, they come with no warranty. Who do you sue if it ruins your equipment or explodes, melts, or catches fire? Spend the extra money on reputable brands. --209.203.125.162 (talk) 00:37, 12 April 2014 (UTC)

Background pictures for Windows 8 desktop

Years ago, when you bought a new computer, it came with several "cool" pictures/designs that you could use for your background on your desktop. Now (Windows 8), it seems that they only give you a very few dull and boring options. Is there any website that provides a variety of backgrounds compatible with Windows 8? Thanks. Joseph A. Spadaro (talk) 23:47, 10 April 2014 (UTC)

Right-click on the desktop and select "personalize". There are themes in there - a collection of related photos. Also, there is a link to get more themes. Bubba73 ^{You talkin' to me?} 00:02, 11 April 2014 (UTC)

Well, exactly. Right-click and "personalize" leads to the 4 or 5 dull and boring themes and photos of which I was aware. I never noticed that link for more themes, though. That looks promising. Thanks. Joseph A. Spadaro (talk) 00:19, 11 April 2014 (UTC)

And you can set any picture from Wikipedia:Featured_pictures as background. Download it and follow Bubba's instructions. OsmanRF34 (talk) 00:08, 11 April 2014 (UTC)

Thanks. I didn't know about that. Usually, though, when I download a random photo like that, it does not seem to quite "fit" correctly on my screen (too big, too small, too narrow, or too "spread out"). Joseph A. Spadaro (talk) 00:22, 11 April 2014 (UTC)

Try this link. Bubba73 ^{You talkin' to me?} 00:58, 11 April 2014 (UTC)

You may want to check your screen resolution at whatismyscreenresolution.com and find a good picture of that size or at least the same aspect ratio. KonveyorBelt 02:36, 11 April 2014 (UTC)

That website says I'm using 1280x720 but I'm actually using 1920x1080. You can check your screen resolution by right-click on the desktop and select "screen resolution". Bubba73 ^{You talkin' to me?} 03:52, 11 April 2014 (UTC)

Could be a big in the site or your browser accidentally or intentionally misreporting.

Could also be that your browser doesn't support DPI scaling but you're using a high DPI (specifically 150% for this case). The new Windows DPI scaling (since Vista) renders the window as a lower resolution and then upscales it for any application which doesn't report itself in the manifest as supporting DPI scaling. You can disable this behaviour on a per application basis in the compatibility settings.

(Sadly even many years after Vista many applications including browsers still don't support DPI scaling. It's starting to change since the Surface and some other tablets and laptops are starting to have high DPI settings by default but last I checked perhaps about 6 months to a year ago neither Chrome nor Firefox properly supported DPI scaling. In fact, this goes back to XP or earlier. The reason the behaviour changed in Vista was mainly because XP and possibly earlier left it completely up to the application but few did. So with Vista if the application didn't say it could handle it, Windows did it the only way it could.)

Nil Einne (talk) 14:54, 11 April 2014 (UTC)

I do have my fonts scaled up 25%, but it is acting like it is 50%. Bubba73 ^{You talkin' to me?} 18:36, 11 April 2014 (UTC)

Usually the first thing I do on a Windows Vista/7/8 PC is download Bliss (image) and change the desktop background to it. 71.3.50.250 (talk) 20:05, 12 April 2014 (UTC)