Dr. Mark Humphrys

School of Computing. Dublin City University.

Home      Blog      Teaching      Research      Contact

Online coding site: Ancient Brain

coders   JavaScript worlds


CA170      CA668      CA686

Online AI coding exercises

Project ideas

Discussion of Internet addressing

Email addresses




Case irrelevant.

Any "." to the LHS:

are just cosmetic. They do not represent subdivisions, as "." to the RHS does.

Site can implement any naming scheme it likes on LHS of the @ sign.

Web addresses (URLs)




Path may map 1-1 to actual path on disk.
Or may not:

Case irrelevant in the http: part and the hostname part.

Case may be relevant in the path part only. e.g. UNIX server maps 1-1 with file on disk and respects case. Windows server doesn't.

Is case sensitivity a good thing?

Computing (CA) server

CA server is UNIX but server software implements case-insensitive username:
but not filename:

See Case sensitivity and "404 Not Found"

Internet addresses v. postal / phone addresses

Internet addresses seem to have a number of advantages over postal addresses or telephone numbers.

Internet addressing is:

  1. Alphanumeric, allowing direct maps to English.
    As opposed to numeric-only addresses like phone numbers.
    (An ad says something like: "14172 - Your new mobile directory enquiries number. The only number you'll ever need." How is anybody to remember something like that? Especially when it changes every couple of years.)

    Be careful about imposing all-numeric or other cryptic solutions on your users. When I was a postgrad, I was given the userid   mh10006   - despite the fact that the userid   humphrys   was unused and available. Why increase login failures, failed emails, and 404 Not Found web requests for no reason?

    It is interesting to consider that underneath, Internet addressing is actually numeric, but machines have been forced to map text strings to numbers since the earliest days of the Internet.

  2. Guessable. I know IBM's web site is ibm.com. I have no idea what their phone number is, and if told I wouldn't remember anyway.
    I don't even know what DCU's phone number is, and I work for them.

    Maybe I could remember IBM's phone number if it was 1-800-IBM-CORP. Or does that 1-800 number only work in America? Or should it be 1-800-BUY-IBM? Who knows.
    And what is their postal address? I don't know, but I know I'll find all that information at ibm.com.

  3. Variable-length. Your company's website is simply company.com, whether that means   bt.com   or   arthurandersen.com.
    It doesn't have to be either squashed into a small number of characters, or expanded to fit a fixed-size field, as with say 1-800 numbers.
    Note that some older competing naming schemes weren't as flexible, e.g. BITNET had maximum 8 char names.

  4. Location independent, so don't have to memorise loads of area codes.
    You bookmark   www.site.co.uk/subsite   and the bookmark works, whether you are in the UK, in Ireland, or anywhere at all in the world. Note that you couldn't set up links for other people if it wasn't location independent.
    Quick! What's the area code for Belfast? What's the area code for London?

  5. Hierarchical, so we can get to dcu.ie safely and then start looking for the Computer Dept. Addresses often contain a self-description of their place in the hierarchy:   comp-dept.western-university.edu.au/staff/dr.smith/phd/references.html  

  6. We can also get repeated feedback in our guesses,
    e.g.   dcu.ie/user   failed, so we try   dcu.ie
    Now we've got to the dcu.ie web home page safely, we can use this as a base from which to explore further and hopefully find a link to   user
    Dial a wrong number and what's your feedback? Post a letter to the wrong address and what's your feedback?

Bad - UNIX usernames

Historically, one area where UNIX/Linux fell down was on usernames.
Long and variable-length filenames were ok, but usernames were restricted to 8 chars.

The legacy of this is still with us. e.g. The UNIX and Linux systems here actually allow long usernames, but the log files and other administration tools only display the first 8 chars - which is why student usernames are still ugly and squashed.

Making the phone system like the Internet

Can we make the phone system like the Web?
We want a phone number system where we can make location-independent links to calls.

If you have a smartphone, or a device with pay-to-call VoIP like Skype, you can do this.



Put the following in your HTML:



Automatic number recognition

Maybe no need to have page authors make the number a link. Do automatic number recognition instead.
  1. Browse page on phone. Select a number on a page. Recognition software tries to dial it.

  2. On PC, Skype Toolbar automatically makes clickable any phone number seen on screen in browser:

    Note how it does not highlight the fax no. Clever.

Old - WAP links

You could set up the following links in the old mobile phone protocol WAP. (Though few did, because few people browsed the web on WAP.)

There is a location independent tag WTAI (also here) in WAP.

Future - DNS system for phones

Might be nice to get rid of (or reduce) the numeric part of the phone numbers:




Would require the implementation of a name server (DNS) system for phones.

Downside: Need to print two different forms of the phone number on your page. One for callers with access to DNS. One for callers with phone-only.

ancientbrain.com      w2mind.org      humphrysfamilytree.com

On the Internet since 1987.

Wikipedia: Sometimes I link to Wikipedia. I have written something In defence of Wikipedia. It is often a useful starting point but you cannot trust it. Linking to it is like linking to a Google search. A starting point, not a destination. I automatically highlight in red all links to Wikipedia and Google search and other possibly-unreliable user-generated content.