Inspired by this Digital Nomads posting:

6 Tips For A Smooth Website Handover

One last point. If you do find yourself sacked by a client for whatever reason, try to play the bigger person and assist the new web developers where you can (don’t go out of your way to help them but give them what they need).

This makes a good point.  Many developers I’ve worked with just get the job done, without respect for other things the client needs.  I’m currently taking over a project where the previous developer was “let go”.  I am in dire need of important information to do the job efficiently and effectively.

And consider supplying a handover check list to the client which lists the information that is required to manage/administer the site.

Your clients are your best advertisers.  The way you treat them is a testament to the quality of your work and to your work ethics.  Even in the best of times, clients decide to change contractors.  Nevertheless, every client should be provided with a checklist of intellectual property items.  Although some things need to shared with the developer, this intellectual property belongs to the client.

The Client is King

The client is the king, and is the only one who should hold the keys to the kingdom.  The most important things the client should have are service provider contact information, account names, and passwords. Even so, there is other information the client should have that will help them move on to another developer, if they find it necessary.

Domain Name

In some cases, the client already owns a domain name.  If they do not, then the client should be guided through the process of creating an account with a registrar, and purchasing the domain name.  Any FTP user name and password that the developer may need should also be created by the client, and you should explain in simple terms what security rights you will need to do your job.  For anyone who doesn’t have a domain name, and would like to do a partial search, I recommend psychicwhois.com.  Most domain name look up services will buy the name out from under you if you don’t buy it right away.  With psychicwhois, you can type in the first few (or more) letters until you see that what you want is NOT taken.  Play with it some time and you’ll see what I mean.

Hosting

Choosing a web host is beyond the scope of this article, but this should be handled in the same way as the domain name purchase. The client should be involved in the entire account creation process.

Database

Once again, there are many different database engines from which to choose, and each has its own user name, password, and host. The database usually doesn’t even reside on the same server as the web site itself, so it’s imperative that the database location, type, and account information be provided to the client at the time of creation.  The client needs to know not only the admin credentials, but also the credentials used by the web application itself.  I recommend two user/passwords for web applications; one had read only privileges, and the other has full read/write/create/drop table priviledges.

Platform

Whether the developer uses Classic ASP, ASP.NET, Ruby, PHP, Python, or Perl, this information should be stated clearly to the client. This will be essential if/when the project is passed on to another developer. If the developer is using a Content Management System (CMS), then the client should be provided with contact information for the company or support group.

Job Posting

Yes, I said job posting.  If your client decides to find another developer, help them specify the requirements that the new developer should meet:

Web site developer needed.  Extensive experience with MySQL, PHP, Javascript, HTML(5), and CSS2 required.

That’s simple, eh?

It’s a little bit me; it’s a little bit you

Now it’s their turn.  When the project is completed, or, if there is an amicable separation between you and the client, then ask them for a letter of reference.  Most satisfied clients are happy to write one, and it’s just one more thing to put into your portfolio.

If you have any other suggestions for this check list, stuff them right into the comment box below.  Thanks

 

Call a URL an EARL

“URL” stands for Uniform Resource Locator, and since I’m so lazy, I pronounce it as “Earl”. It saves you 7 syllables, and possible embarrassment if you accidentally said “uniform refund lecturer”. Also, being from the USA, when speaking, I would say “an EARL” rather than “a EARL” since EARL begins with a vowel. If I were going to actually say the letters separately, I would say ” a U R L”, instead of “an U R L”.   If i am typing it in an email message, or, this post, for example, I would use “an URL”, while mentally hearing “an Earl”because I type pretty much like I speak.

Testing servers

We coders and/or network managers use multiple protocols to do our jobs.  There are web servers, ftp servers, mail servers, and database servers.  You can at least test them to see if they’re up by using telnet, and knowing a few commands can help you test functionality.

HTTP

Web servers default to use port 80.  They can be set to listen to other ports, but 80 is the default.  If you set one up to run on port 8000, you would have to use  http://www.example.com:8000 to get there, AND the scheme HTTP:// is required, because browsers can handle multiple schemes, which designate the protocol.  Think of a protocol as a dialect of English.  Using anything other than the default port is required to let the browser know to use the http protocol.

So, let’s test the web server to see it it’s up.  Make sure that you’re at the Linux or Windows command line:

telnet www.microsoft.com 80

If it’s not able to connect, you’ll get an error message, but in this case, we all know that Microsoft should always be up (?).  You should see a message saying that it is connecting, and then the screen will go blank.  Now test the functionality of it:

GET /

Press ENTER twice.  The first ENTER ends the GET command.  The second ENTER tells the server that you are through with the commands.  You should see some html there, like at least a <html><head><title>Microsoft Corporation</title> all the way down to the </html> tag.

The web server is up and functioning.  Once you receive the HTML, it will close the connection, because HTTP is a stateless protocol.  You make a connection, make a request, get the results, and the connection is closed.

FTP

FTP is an entirely different protocol, and uses two different ports. The default ports are 21 for the commands, and 20 to transfer the data. We’ll only test the command port here.

telnet www.microsoft.com 21

Telnet to ftp.microsoft.com on port 21, and type “HELP” and hit enter. You’ll see a listing of all valid commands that their server supports.

220 Microsoft FTP Service
help
214-The following commands are recognized (* ==>’s unimplemented).
ABOR
ACCT
ADAT *
ALLO
APPE
….

So, the ftp server is up and functional. Whether or not you are authorized to get or put files is a matter of authentication.

SMTP

This is Simple Mail Transport Protocol, and is only used for sending email.  The default  port is 25.  Talking to your mail server is just as easy as HTTP and FTP.  Telnet to mysmtpserver.com on port 25.  You should see a 220 greeting message like this (all lines with a numeric code at the beginning of the line came from the server):

220 Welcome Road Runner. WARNING: *** FOR AUTHORIZED USE ONLY! ***
ehlo austin.rr.com
250-hrndva-omtalb.mail.rr.com says EHLO to 70.114.203.216:56199

To send a message, use the commands shown below.  :

mail from:<squeek@austin.rr.com>
250 MAIL FROM accepted
rcpt to:<squeek@austin.rr.com>
250 RCPT TO accepted
data
354 continue.  finished with “\r\n.\r\n”
this is a test message
hit return, then a period, then a return to end the message, as you see above (\r\n.\r\n)

I can hit enter as many times as I want to

but enter, then . then enter, and I’m done.

.

In this case, my server refused the message, because I didn’t actually log in using my username/password.  We wouldn’t want to let everybody and their chihuahua send SPAM via this server, correct?

POP

This is Post Office Protocol, and is only used for receiving email. The default  port is 110.  Talking to your POP server is just as easy as talking to your SMTP server, except the protocol (commands) are different.  Telnet to mypopserver.com on port 110.  You should see a 220 greeting message like this (all lines with a numeric code at the beginning of the line came from the server):

Databases

We can do the same with databases, but the actual commands used are pretty complex, so I won’t go into that.  The default port for Microsoft SQL server is 1433, and for MySQL it’s 3306.

The server that I have MSSQL Server on won’t come up.  It worked the last time I turned it on, but this time it didn’t.  That’s when things usually break, folks – when you turn them on.

OK, for MySQL:

telnet 127.0.0.1 3306

You should get a bit of garbage, but you should at least see the version of MySQL you’re using.  I get 5.5.16, which is correct.

Ports

Most of us in the business already know this stuff, but for the others…
“So what are all of these port numbers” you ask?  Think of them as doors into your house (server).  Webb (the web server, and my father’s first name) only answers the front door, one request, (Can Billy come out to play?), answers it (No, he’s got the chicken pox), and promptly closes the door (as we saw above when testing – the connection is immediately closed).

Freddie (FTP) only listens to the back door, waits for a request, and (hopefully) sends you a file.  He doesn’t close the door like Harold does, because you might want to ask for another file.

Sammy (SMTP mail server) only answers the side door.  If you show him your identification, he will then accept a message from you to be sent to someone else.  Like Freddie, he leaves the door open, in case you want to send another message.

To repeat: Web servers default to port 80.  FTP servers default to port 21.  SMTP servers default to port 25.   A server can have multiple web servers, web sites, ftp services, and smtp servers; all you have to do is use different port numbers on each service.  Universities used to use multiple ports on their one web server to differentiate between the different departments.  They may still do that; I’m not in college any more.

Keep in mind that if you add too many doors, the house (server) will eventually fall down.