Wikipedia:Reference desk/Archives/Computing/2017 February 18

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

February 18

Website without domain name

(1) If I want to have my own website, can I reserve a numerical address (IPv4 or IPv6) without cost, and thereby avoid the cost of registering a domain name? (2) If I am obligated to accept an IPv6 address because no IPv4 addresses are available, can I promote the website by submitting the website name to search engines?
—Wavelength (talk) 00:21, 18 February 2017 (UTC)

It depends a lot on how your web content is hosted. Do you already have a host set up? clpo13_(talk) 01:07, 18 February 2017 (UTC)

No, I do not. What do you recommend in web hosting?

—Wavelength (talk) 01:11, 18 February 2017 (UTC)

(1) No one will use your site without a domain name. Also if/when your IP address changes everything with the old address on it is now wrong. It is technically possible to connect with just an IP address, but a) there is no guarantee every piece of software will allow this, and it might be outright prevented by some institutional (corporate, school, etc.) networks; b) average people will wonder what the heck that string of alphanumeric gibberish is. Domain name registrations are not very expensive. (2) You don't "submit" your website to search engines. That's an idea from the good old days of '90s "Web portals" like (the original) Yahoo and Lycos, where actual humans compiled listings of sites. Today's search engines work by automatically crawling the Web, following hyperlinks. To get your site to show up in search engines, you have to get other people to link to it. --47.138.163.230 (talk) 03:32, 18 February 2017 (UTC)

Pair.com ( https://www.pair.com/webhosting/shared.html ) is good. $5.95 per month basic hosting. Using an IP address is a Bad Idea, because dedicated IP addresses cost more than domain names. Shared hosting (many domain names on one shared IP address) is the way to go. Namecheap ( https://www.namecheap.com/ ) is a good place to buy a domain name. --Guy Macon (talk) 03:41, 18 February 2017 (UTC)

Ah the good old days! When one did one's own routing typing in the various ip numbers to get there. What is the point of a domain nobody ever looks at? Because that's what you'd have. And it would be more expensive too. Dmcq (talk) 09:57, 18 February 2017 (UTC)

Without a domain name, you're going to have trouble finding cheap hosting. They want to put a bunch of virtual servers on a smaller number of physical servers. That presents difficulties without a domain name. (Although some might allow you to use a subdomain off of the hosting company's domain. That used to be common back when domains were expensive. Haven't seen it in a while, though.)

If you're setting up your own physical server in your basement, it might be feasable to save $9.95/year by skipping a domain, but only if your ISP offers static IP addresses, which they probably don't.

Nowadays there's not much point to trying to avoid buying a domain name. They cost only ten bucks a year for a dot-com and some of the weird ones cost even less. dot-party domains seem to be available for a buck a year. What a deal! ApLundell (talk) 14:49, 20 February 2017 (UTC)

Reporting spam/phishing

I've recently received some emails that looked like spam or phishing, and since they were sent from addresses affiliated with major universities, I notified both institutions' IT security offices about these incidents. However, in both cases, my reports got rejected immediately by the recipient addresses (apparently they thought I was spamming or phishing them), so I had to report these cases in a more roundabout manner. The same thing has happened at other times in the past — I can't report something somewhere because the report is misinterpreted by the receiving email server, or even I can't send the report because my email server thinks that my computer's been taken over by a botnet and sending out spam itself.

Since computer science is constantly studying ways to make computers act like humans (e.g. passing the Turing test) and ways to ensure better information security, and since tons of institutions maintain abuse lines where people can report fake emails, I'm guessing that scholars have studied the fact that accurate reports of disruptive emails can get misidentified as disruptive themselves. Could someone point me to research on the subject? Nyttend (talk) 06:20, 18 February 2017 (UTC)

Sorry, this isn't what you're asking about, but you do realize that the institutions aren't actually involved, right? While the emails might have looked like they came from Harvard and Yale or whatever, they actually came from some anonymous creep. What you did was right-intentioned, but there's not a lot they can do about it. Mark as spam, delete, and move on. Matt Deres (talk) 13:11, 18 February 2017 (UTC)

As Matt writes above, it's probably best not to pass on any e-mails with suspicious-looking attachments because they probably came from a dangerous source and had a spoofed sender address. The universities probably have better spam filters than you have. The exception is suspicious messages purporting to come from banks. These organisations often ask for fake messages to be forwarded to their phishing department so that they can take appropriate action to take down any fake websites. I don't know of any research reports. Perhaps someone else can link to some? Dbfirs 13:35, 18 February 2017 (UTC)

• Then there's the other problem with reporting it to a bank. --76.71.6.254 (talk) 23:03, 18 February 2017 (UTC)

With both emails, the sender's "name" was that of a different university, but in both cases, the actual addresses were university-affiliated, and when I decided to get around the spam filter by +poning people with the IT help desk, they actively encouraged me to forward these emails. Nyttend (talk) 15:18, 18 February 2017 (UTC)

What do you mean by "the actual addresses"? It's pretty trivial to spoof most of the headers in an email, so the From, Reply-To and similar headers can't be trusted. The only things that can really be trusted are those that are cryptographically secured, like DKIM-Signature. CodeTalker (talk) 18:09, 18 February 2017 (UTC)

You can trust the last entry in the path -- that's your email system. And you can trust the next one up -- that's where your email system got the email from. If that system is one that you trust, then you can trust where it says it got the email from, and so forth. --Guy Macon (talk) 18:53, 18 February 2017 (UTC)

You're referring to the Received headers, and what you say is true, as far as it goes. But Received headers can be forged too, and many legitimate emails pass through systems which you may not have any knowledge of, so can't tell whether you should trust them or not. CodeTalker (talk) 19:57, 18 February 2017 (UTC)

But the fact they can be forged is largely irrelevant since what can't be stopped is your own email system's practice. If you know your own email system's practice you should know which received header was placed by it. It's fairly unlikely your own email system's header was forged anyway, it's just too complicated for most spammers so you don't even need to know where exactly your email system will place the header, just which one it is. But in any case, you just need to pay attention, if you're not sure where your email system will place the received header, just make sure there is only one. If there seems to be 2 each claiming the email was receive from 2 different places then you know there's a problem.

While you're right you can't always be sure of whether you can trust the system the email passed through, often you can be resonably sure for emails from large businesses since they use well established systems and nowadays emails tend to pass through very few external hops. The key point here that I often make (and I think may be at least partially Guy Macon may be making) is that while people often claim you can't tell from the headers whether the email actually came from your bank (or whatever), by the headers without DKIM etc, because everything can be forged, this often isn't true.

If you know what you're doing (and to be fair, this is something beyond the average user) you can generally get a good idea. If your email system says it received the email from your banks email systems (from my experience this is fairly common with important emails send by the bank), then it's fairly likely the email was sent by someone in the bank, or your bank has major security flaws and are allowing unconnected parties to use their system to send emails.

Of course you probably shouldn't always trust everything sent by your bank. And I also question the security practices of a bank who don't use SPF etc nowadays anyway. But the point is these concerns arise for other reasons not because you can't work out from the headers whether it's likely your bank sent the email. Now if your email system says they received it from spammer.spam or some random small business, then it's fair to think it probably wasn't sent by your bank.

If spammer.spam or the small business claims they received it from your bank then yes that's probably forged. Although with modern email practices you don't even really need to know your banks norms to guess there's something dodgy the moment you see your email system received it from spammer.spam or some random small business.

I'm not denying there can sometimes be uncertainity. For example, one bank I referred to which sends important emails directly from their email system also uses Campaign Monitor for at least some of their advertising campaigns. Is Campaign Monitor trustworthy enough that if they say it was coming from my bank this is true? Probably but depending on your security requirements this may not be enough.

Nil Einne (talk) 04:07, 19 February 2017 (UTC)

While I agree with others from your description most likely the uni wasn't actually involved in the emails, there seems to be some contradiction here. It's not clear from your original comment whether this was a phising attempt, but there's no reason why only a bank is going to want to shut down websites used for phising. Any large organisation and most small ones are likely to want to avoid their customers (or whatever) authentication info being stolen. The question is more whether the organisation can do anything than the particular organisation. Banks do probably have larger and more dedicated teams, but that's ultimate the key difference. If someone is sending malware, whether it's claiming to be from a bank, a university or FedEx, most likely none can do much except warn their customers if necessary. This may be more likely with the bank although some other organisations may be just as likely to warn. If someone is sending a 411 scam or adverts for "Viagra", again there may be not much that will be done whether it's purported to come from a bank, a university or FedEx. As said, if someone is attempting to phish using external websites all organisations are probably going to want to shut those down. It may be more likely that something purported to be a bank is a phising attempt but something claiming to be from a university is something else, but that's a different point. Nil Einne (talk) 04:22, 19 February 2017 (UTC)

If you receive spam or phishing emails, per RFC 2142, you should be able to forward them to an abuse email address. E.g. spam from something@yale.edu can be forwarded to abuse@yale.edu. Ideally, your email won't be rejected, as the mail server should not be screening emails to abuse. That's the whole purpose of that particular email address - it accepts fishy emails to be reported. Spam reporting has some limited info. ^Avic_ennasis @ 18:24, 23 Shevat 5777 / 18:24, 19 February 2017 (UTC)

CPU question

If you took out the CPU chip out from the motherboard and carefully soldered wires from each contact pin to the corresponding pin on the motherboard, would the CPU function? Would it be slower than directly connected? — Preceding unsigned comment added by 189.201.241.54 (talk) 19:28, 18 February 2017 (UTC)

I did something alike in 1980 with an RCA 1802 8-bit CPU and it worked. I believe the clock speed was 1 MHz. Today CPU's are much faster (clock speeds more than 2000 times faster). High speed electronics needs a careful PCB design, not sloppy wires. It seems highly unlikely to me that it would work without drastically reducing clock speed (which could cause it's own problems). Jahoe (talk) 20:50, 18 February 2017 (UTC)

Yes, the signals take about a nanosecond to travel a foot, so if you have several inches of wire the delays would almost certainly cause timing problems for gigahertz clock speeds. Dbfirs 21:04, 18 February 2017 (UTC)

Good point to mention the delays. I thought in terms of impedance mismatch, signal reflections, parasite capacity, etc. In other words, signal spaghetti. ;) Jahoe (talk) 21:40, 18 February 2017 (UTC)

There's also capacitance between the wires to consider. Several inches of wire would create extra capacitance of maybe 20-50 picofarads between them. This capacitance would have a slowing effect on the rise time of the signals the wires carry, meaning that the signals would not rise or fall as rapidly between zero and full voltage states as quickly as they did previously. Unexpected behavior of the CPU would occur in time-critical operations. The extra capacitance also means that a false signal might be induced to appear in an adjacent wire, creating havoc. In general, the computer could be expected to run slow and probably crash. Akld guy (talk) 19:22, 20 February 2017 (UTC)

what were early 8-bit home computers (with a front panel) good for?

I'm thinking of the Altair and the IMSAI Asmrulz (talk) 21:49, 18 February 2017 (UTC)

The same thing other computers of the era were good for. Since you mentioned the front panel, I suspect you're thinking all people did was plug them in and watch the lights blink. Actually, to do useful things, you connected them to peripherals. The Altair article mentions connecting it to other devices like teletypes, often via RS-232. A lot of companies were started to sell hardware for the new microcomputers. This was no different from the mainframes and minicomputers of the era. For instance, here are Dennis Ritchie and Ken Thompson at work on the Bell Labs PDP-11 sometime in the '70s. --47.138.163.230 (talk) 22:06, 18 February 2017 (UTC)

I'm not thinking that. But you have to admit the I/O capabilities were limited. And how many people had a discarded teletype sitting around, anyway? I've been toying with microcontrollers for the past year or so. Hence I'm curious what people typically used computers that are hardly more than a glorified CPU evaluation board for. Asmrulz (talk) 22:53, 18 February 2017 (UTC)

The two types you mention weren't really home computers (a bit too expensive for that), but machines for small businesses. They were used for financial administration, word processing, inventory, spreadsheets, etc.

Home computing didn't really take off before the 1980s (3 to 5 years later than the types you mention). These home computers were used for games (tetris, kings quest, leisure suit larry, etc.), word processing (WordPerfect) and as a general introduction to the new technology. Many owners stopped using them after a while (stuck with the same question as you ;), others (like me) got enthusiastic and went on a quest for a real purpose of their new and beloved toy. ;)

Sometimes (varying per country) employers financed home computers to have their staff practice at home in their spare time. (Doesn't sound good perhaps, but I loved it.)

In my view, home computers were fun, but didn't have a real purpose until the internet came to the family home in the 1990s. And even today, most PCs at home are little more than internet terminals. They're a bit like telephones, which have little use without a telephone network.

Well, that's how I look back on it from my own experience. Other people may have completely different thoughts on it. But perhaps this is a (partial?!) answer. :) Jahoe (talk) 22:58, 18 February 2017 (UTC)

TeleVideo 925 glass teletype

Oh, and those computers with front panel weren't typically used with teletypes, but with "glass tty's", like the Televideo pictured. Connecting two or four terminals was possible, although one was more common. Jahoe (talk) 23:09, 18 February 2017 (UTC)

The Altair and IMSAI weren't really marketed to businesses; they were marketed to nerds who wanted to mess around with computers (and I say that as a proud nerd). An individual (with sufficient disposable income) could buy their very own computer to tinker with, some assembly required. That was a big deal when up to that time computers took up at least a room, cost as much as a house, and required at least several experts for care and feeding. The first microcomputers that were marketed to businesses and professionals for "serious work" were the Apple II and IBM PC (which, of course, didn't require you to assemble them from parts). See also history of the personal computer. --47.138.163.230 (talk) 00:18, 19 February 2017 (UTC)

Although the Altair and IMSAI were very popular, there were other early computers. I myself started with an NCR Century 100 (My high school bought one and started teaching classes on computers with it), and my first home computer was a COSMAC VIP, which I built from a kit in 1978.

You too can own a COSMAC ELF! See [1] and [2]. --Guy Macon (talk) 01:25, 19 February 2017 (UTC)

I was around at the time. They were so limited that I didn't want one. I either used a mainframe or my programmable calculator, which could store programs to a magnetic strip and read them back. I also had a printer to use with it. I think that a programmable calculator at the time was a lot more useful. Bubba73 ^{You talkin' to me?} 01:42, 20 February 2017 (UTC)

DEC GT40 graphics terminal running Moonlander

They where special "adult toys", like model railways. You could do little basic stuff with it but they where very fascinating for everyone who liked Enterprise/Star Trek: The Original Series. Actually there also where some very early, simple games availabel back then, like Lunar Lander. They where very rare tho as maybe 1 out of 10,000 People, even or less, had one, which made it double fascinating ofcourse, so they where priceless, for bragging about having one, for some. --Kharon (talk) 02:14, 20 February 2017 (UTC)

Radio amateurs began using early 8-bit computers for sending Packet radio messages, keyboard to keyboard. The first continent-to-continent packet radio conversation was made quite a few decades ago. However, most operation was, and still is, on the 144 and 432 MHz amateur bands. Radio amateurs also set up Bulletin board systems, interlinked between nodes in major cities via radio a decade before the internet took off. Akld guy (talk) 04:32, 19 February 2017 (UTC)

interesting. where I'm from, non-IP, modem networks such as FIDO were in use well into the late 90's-mid 200x (landline, not radio, though) Asmrulz (talk) 04:59, 19 February 2017 (UTC)

First person to send fax and slowscan TV over the cellular network

I believe I was the first person to achieve this. How can I place this in Wikipedia in order to determine if this is the case?Bob Moore 100 (talk) 22:29, 18 February 2017 (UTC)

You can't put WP:original research in any Wikipedia article, but if details of your achievement have been published elsewhere in WP:reliable sources then the fact can be recorded here. It is not possible to use Wikipedia to determine whether the claim is true or false, though someone might be able to find references to similar claims. Dbfirs 22:45, 18 February 2017 (UTC)

Also, on Wikipedia it's considered not done to write about yourself or your own achievements (at least not outside your user page). Of course others could do that for you just fine. Jahoe (talk) 23:34, 18 February 2017 (UTC)