Welcome to the computing section
of the Wikipedia reference desk.

Select a section:

WP:RD/C

Want a faster answer?

Main page: Help searching Wikipedia

How can I get my question answered?

Select the section of the desk that best fits the general topic of your question (see the navigation column to the right).
Post your question to only one section, providing a short header that gives the topic of your question.
Type '~~~~' (that is, four tilde characters) at the end – this signs and dates your contribution so we know who wrote what and when.
Don't post personal contact information – it will be removed. Any answers will be provided here.
Please be as specific as possible, and include all relevant context – the usefulness of answers may depend on the context.
Note:
- We don't answer (and may remove) questions that require medical diagnosis or legal advice.
- We don't answer requests for opinions, predictions or debate.
- We don't do your homework for you, though we'll help you past the stuck point.
- We don't conduct original research or provide a free source of ideas, but we'll help you find information you need.

Ready? Ask a new question!

How do I answer a question?

Main page: Wikipedia:Reference desk/Guidelines

The best answers address the question directly, and back up facts with wikilinks and links to sources. Do not edit others' comments and do not give any medical or legal advice.

October 9

Coaxial cable

I remember in the days of analog satellite I used to have a l "coaxial cable" that specially connected the satellite receiver and TV and served as a way to directly send the sat signal to the TV so that it would appear as a fake channel instead of having to use the AV button.

Some time ago I had digital satellite set up and I've been told I also need a "coaxial cable" to send the signal to the TV so that I can watch the free unencrypted channels with the TV off, or route signals from another dish as well. Or something like that, the free channels are important because now I can either plug the antenna cable into the TV and watch only free channels, or plug it into the receiver and see everything but only on receiver (I have to watch the local channels also via receiver).

I've duckducked coaxial cable but it seems that's something else entirely. What's the thing I need called? 95.168.116.4 (talk) 14:59, 9 October 2019 (UTC)[reply]

Analog and digital are converted back and forth on various devices, so coax cable can still be part of the equation, and has the advantage that nearly every TV can accept it, even very old ones. There are other TV input options, but not every TV has those. And it sounds like you need a switch box [1], which allows you to press a button to swap which coax cable is plugged into the TV. (For some reason, TVs rarely have more than one coax input, which would allow you to use the remote to switch between them.) Another option is to use a splitter [2], so both coax cables are plugged in at once. This works if the two inputs use different channels, but expect some degradation in signal quality over using the switch box. Still, if the degradation is minimal, this might be better than having to get up to hit the button to switch inputs (but you'll likely still need to switch between cable and antenna inputs on the TV menu). SinisterLefty (talk) 16:20, 9 October 2019 (UTC)[reply]

Ah, of course, the splitter. We had that on one TV. On the other one the sat reciever had two cable jacks. One had the antenna cable and the other had the cable that the guy called coaxial running from the reciever to the TV. The digital receiver also has two similar looking cable jacks, although the antenna cable only goes in one of them, so this probably needs another type of cable if it's for that at all. The other one is what got me thinking about this cable, I forgot splitters were a thing. 95.168.120.97 (talk) 22:31, 9 October 2019 (UTC)[reply]

Yes, some set-top boxes have a bypass, that allows you to pass whatever comes in (on the coax) back out to the TV without doing anything to it. If you have that, then you can feed the antenna signal into that set-top box, and use it's remote to control whether you see the antenna signal (the bypass) or let the set-top box do it's thing and unscramble the satellite signal, etc., and send that to the TV. (Of course, with modern thin TVs, the "set-top box" isn't actually on top of the TV set anymore.) SinisterLefty (talk) 02:59, 10 October 2019 (UTC)[reply]

Thanks, I think the splitter is likely what I'm after. 95.168.116.28 (talk) 21:56, 10 October 2019 (UTC)[reply]

As to the switch box, you may also find it under the name "A/B switch". --76.69.116.4 (talk) 03:27, 10 October 2019 (UTC)[reply]

Yes, for a 2-way switch box, but they also make 3-way, 4-way, etc. SinisterLefty (talk) 03:50, 10 October 2019 (UTC)[reply]

October 10

October 11

Apple Store VS Bestbuy for Repairs?

I Live in California, United States and will be sending in my iPhone 6 for battery-related repair. Does anyone know if there is any difference in pricing or customer satisfaction between the Apple Retail Store and BestBuy? 70.95.44.93 (talk) 12:32, 11 October 2019 (UTC)[reply]

I've never taken anything that wasn't under warranty for repair to a chain store, but any decent repair shop should be able to open it up for free or a small fee and give you a quote for the repair cost. 93.136.47.32 (talk) 15:10, 11 October 2019 (UTC)[reply]

If the "battery-related repair" is simply replacing the battery (in a phone where that can't be done by the user), then both should be able to quote you a price, up front. SinisterLefty (talk) 04:19, 12 October 2019 (UTC)[reply]

Yeah battery replacement in an iphone requires a special tool but is otherwise pretty easy. If you want to do it yourself, look on ifixit.com. They sell a replacement kit that includes the tool and a new battery. 173.228.123.207 (talk) 22:46, 15 October 2019 (UTC)[reply]

What is Wiki text?

I recall there is a computer term for Wikitext (or wikisource) meaning that it is semi-structured. It has some structures (section headers, citation templates etc), but it's also not so structured as a database like Wikidata. It's somewhere between free-form plain text with no real structure and a database that is completely structured. What is this type called? -- GreenC 14:37, 11 October 2019 (UTC)[reply]

There can be many terms, such as "markup language", which is a language in which markup is included in the language. 135.84.167.41 (talk) 16:16, 11 October 2019 (UTC)[reply]

Note that WYSIWYG ("wizzy-wig") is a term for the plain text. Here's a link to markup language, with HTML being one of the most popular. SinisterLefty (talk) 17:22, 11 October 2019 (UTC)[reply]

Thanks! There is a more general term/concept I'm looking for. I've seen it before but can't remember but those are not it. -- GreenC 17:26, 11 October 2019 (UTC)[reply]

Rich text? 93.136.47.32 (talk) 18:00, 11 October 2019 (UTC)[reply]

It has to do with semi-structured vs. structured. -- GreenC 20:42, 11 October 2019 (UTC)[reply]

Also see wiki, which talks about the rich-text editor mentioned above. SinisterLefty (talk) 04:21, 12 October 2019 (UTC)[reply]

I think "markup language" is about as good a term as there is. Semi-structured is ok too, but it doesn't "officially" mean anything. 67.164.113.165 (talk) 16:59, 13 October 2019 (UTC)[reply]

Backlink request

My name is Lysa Ward,MD and I'm an expert provider of the Fraxel 1550, also known as the Fraxel Re:Store treatment. My practice URL is [3] and I would like to request a backlink. The official page on Wikipedia for the Fraxel treatment https://en.wikipedia.org/wiki/Fraxel is somewhat short. I have 20 years of experience and I think the reader will have several unanswered questions still. Thank you very much for your time and consideration.

Sincerely,

Lysa Ward, MD

SkintegrityMedSpa (talk) 21:54, 11 October 2019 (UTC) SkintegrityMedSpa (talk) 21:58, 11 October 2019 (UTC)[reply]

You can ask for this at Talk:Fraxel. -- Hoary (talk) 22:11, 11 October 2019 (UTC)[reply]

Or not, because OP's clearly a self-promotional spam account. Ian.thomson (talk) 23:38, 11 October 2019 (UTC)[reply]

I'd put it a bit differently. The reason we shouldn't link to that page is that the page is promotional, not that the OP is. Even if somebody with no link to that company whatsoever wanted us to link to that page, it would still be inappropriate. SinisterLefty (talk) 06:32, 12 October 2019 (UTC)[reply]

October 13

set difference in SQL

Again this is part of my "learn SQL" mission so I'm looking for the "best" (scaleable etc.) answer rather than just a usable one. I'm trying to get news updates by reading a news site once an hour or so, extracting the links to news articles, and checking which links weren't there before (news stories tend to stay on the front page for many days). In sqlite, "N+1 queries are not a problem",[4] so it's ok to just store the old urls in an indexed column, and then when doing an hourly scan, do a separate lookup for each link in the page. But I'm wondering how to avoid this. There can be up to a few hundred links on the page, e.g. on www.reuters.com or www.sfgate.com. My idea is:

Have a permanent table A of old urls, indexed on url
on scanning a page:
- create an unindexed temporary table B in memory and insert all the just-read urls into it
- create an index on the temporary table
- CREATE TEMPORARY TABLE C AS (SELECT url FROM B WHERE url NOT IN A);
- INSERT INTO A FROM C;
The idea above is that the query planner should be able to do the right thing with the above queries to create table C by using the indices on A and B to scan both in order, making a single linear pass, then do the table merge efficiently as well.

Is the above approach ok? What is the best way to create a temporary table with 100s of entries supplied by the application? Does it vary much by db? I'm using Python/sqlite3 atm but want to become reasonably clueful about postgresql and mysql as well. (Don't care about Oracle or MS SQL). Thanks! 67.164.113.165 (talk) 17:20, 13 October 2019 (UTC)[reply]

Assuming that these tables are just for your use, then that sounds like a good way to do it. On the other hand, if others would need to access those tables, then it might be better to break that down to smaller operations to avoid table locking. That is, you could check each new URL link against the table of old URL links, and only insert it if it's not found. Probably less efficient overall, but might handle the table locking better. Now this wouldn't matter for hundreds of URLs, but you asked about scaleability, and if you went to millions or URLs, then it might well matter. Also note that the index should probably be recomputed each hour, after inserts are done.

Also, unrelated to DBs, your method of finding if an update has occurred seems likely to miss some updates, where content was changed, but the URLs in the links were not. SinisterLefty (talk) 02:11, 14 October 2019 (UTC)[reply]

Thanks, yeah, sometimes a story might get updated but it's ok, I can always look on the live site. Yes the actual program is just for my own use, but I'm still looking for how a real db dev would do it. I thought looking up urls one by one was an N+1 query which is what I'm trying to avoid. Like if I were doing this in a Python program instead of sql, I'd use two sets (basically hash tables) so I could quickly do in-memory lookups without any db traffic. The corresponding sql thing would seem to be push all the new urls to a temporary table (maybe not all at once, ok) and then do some bulk operation on the tables.

Do the high end db's like postgres lock entire tables when doing updates? I had hoped not, but I guess it's something to consider. I bought a book about db optimization so I'll see if I can find any suggestions about this. So far I've only skimmed it to get the general ideas, which mostly make sense. 67.164.113.165 (talk) 04:28, 14 October 2019 (UTC)[reply]

If we consider that there were two users, and one was doing a search for a URL, while the other was adding that URL to the table, then yes, it should properly lock the table while doing the update so that the search gets the correct results. However, there might be a way to specify that an occasional "wrong answer" is OK and that therefore it shouldn't lock the table when doing updates. There could also be issues with updating the index. If the index isn't updated, it would then take far longer to search for a URL. So, again it might make sense to just not allow searches until the index has been updated, although this would only matter if thousands of URLs had just been added.

I believe the way some massive DBs handle this is that they hold all updates until the nightly maintenance cycle, and update the index then. This allows for searches without delay during the day, with the understanding that they won't reflect the latest info, which "isn't in the system yet". SinisterLefty (talk) 04:59, 14 October 2019 (UTC)[reply]

Oh, and one other comment on your original plan, I wouldn't create the index for the temporary table. As you said, the DBMS can figure out the best way to do it, and that includes creating an index, if that would help (but I don't think it would). The permanent table, on the other hand, needs the index, as you don't want it recreating a temporary index every time. SinisterLefty (talk) 05:08, 14 October 2019 (UTC)[reply]

Ah, good point, with btree indexes it's enough that only the permanent table has it since the temp table entries will be looked up one by one at the db end and it will still be O(N log N) operations. I had imagined a sequential scan of both indexes side by side (like looking for common elements in two sorted files) but don't know if it would really work that way. If I get really bored I might try doing some timings. Thanks. 67.164.113.165 (talk) 05:24, 14 October 2019 (UTC)[reply]

Yes, bench-marking is certainly the way to find the best option for that particular DB, table implementation, data size, etc. However, if anything changes, even a minor update of the DB system, the results could change dramatically. Also note that you would want to compare the time without an index against the time with an index plus the time to create that index. SinisterLefty (talk) 05:32, 14 October 2019 (UTC)[reply]

Another general comment: This isn't a good way to determine if a web page has changed, unless you want the list of URLs for some other purpose, like a search-engine spider. If not, something like a diff would make more sense. SinisterLefty (talk) 05:56, 14 October 2019 (UTC)[reply]

They move the links around on the page but the links generally point to articles so I think the set difference approach is ok. It's not a general purpose scraper, it's just for a few specific sites, currently reuters.com and sfgate.com. Both of them currently stick to a fairly rigid template where there are sections that each contain a bunch of links with headlines. There are some other sites like nytimes.com where the main page contains story text and where diffs might work better. I hadn't really thought about that, hmm.

Actually what am I doing looking at the HTML at all? I should pull the RSS feeds instead. I wrote an HTML de-junker anticipating the California power outages (so I could check the news with limited mobile bandwidth) and it got a bit out of control. But I should rethink it. Thanks for the ideas. 67.164.113.165 (talk) 17:24, 14 October 2019 (UTC) (Added: Actually I can use RSS to list articles instead of dejunking the front page, but I want have to dejunk the articles themselves, since they are full of videos and crap that drive my browser berserk even during normal times. So I'll keep using the dejunking code, which strips out a lot of scripts and images etc. 67.164.113.165 (talk) 20:39, 14 October 2019 (UTC))[reply]

You're welcome. And can I get a copy of that dejunking code ? Sounds useful ! SinisterLefty (talk) 17:28, 14 October 2019 (UTC)[reply]

Sure, lemme clean it up and figure out how to upload it. It basically just loads the page into BeautifulSoup (html parser, "pip install bs4") and then deletes the JS scripts, images, and some other crap found by hand-examining the page. You get a very fast loading page that has the same general shape as the original. I'll be away for the rest of this week but will be back after that so will try to upload it then. I also like lite.cnn.io which is pretty junk-free to begin with. Also: text.npr.org . I did reuters because lite.cnn.io was offline for a while, and then sfgate because it had the most current updates about the CA power outages. It's fairly simple to do other sites though as long as they don't rely too much on Javascript or AJAX. 67.164.113.165 (talk) 04:41, 15 October 2019 (UTC)[reply]

Thanks, those sites are nice, although I don't mind pics so much, just the ads, and animation or video that plays on it's own. SinisterLefty (talk) 05:49, 15 October 2019 (UTC)[reply]

October 15