It's
the Content Management, Stupid
by Jim Howard
New Media, August 8, 2001
The objective is to store assets -- in this case, digital
assets -- efficiently and then rely on a catalog system for quick and easy
retrieval. One approach to asset management involves developing a set of rules
for manually naming and creating versions of files on your file server. Another
approach is to buy and install a digital asset management system.
Those rules and systems sound like a manager's policy and
purchasing decisions. So I'd suggest a third approach, which will best prepare
you to be a manager making those policy and purchasing decisions: develop your
own rules and your own content management system by building and maintaining a
web site.
Then at the job interview when they ask, "Can you manage
with the Web?", you won't have to gulp and reply, "Well, I took notes
during a lecture once in MBA school." Instead, you'll be able to display
your site on a laptop and demonstrate your decisions about the rules and
policies for your own site.
a note on terminology
It gets confusing when people use "site" to mean every
unit from a page to a home page to a web to a web site.
They say, "Go to that site," when they mean "Go
to that page." They say, "On that site," when the mean "On
that page."
Make sure that we're using web making terms such as site,
web, and page the same way.
W3C's Web
Characterization Terminology
The Web has proceeded for a surprisingly long time without
consistent definitions for concepts which have become part of the common
vernacular, such as "Web site" or "Web page". This can lead
to a great deal of confusion when attempting to develop, interpret, and compare
Web metrics.
Directories
You'll sometime see a directory referred to as a folder.
The icon usually associated with a directory in Windows is a little yellow
folder. I tend to use directory when I'm talking about the server and to use
folder the rest of the time.
Directory structure
Directory structures are characterized as flat (or shallow) or
deep. It's up to you. As long as each of the 500+ pages on Ricci Street had a
discrete file name, they could all be in the same directory. It would have a
very flat or shallow structure.
It would also drive me crazy. Maintaining it would be one long
scroll after another because it's a longer list than can fit on my screen all at
once. Out of necessity, I would try to exploit the alphabet to help me. I'd
probably end up with a code that no one else could read or with very long file
names such as:
gizmostoolkitwebmakingwebsitemanagement.htm
Instead, I group the files into directories for the same reason
I group files on my hard drive. It makes them easier to manage. It lets the file
names be shorter.
Thus, I have made a deeper structure to make my life easier. It
also eases the life of an experienced user of Ricci Street. For example, you
could be at this page
http://RicciStreet.net/dwares/lane/mba600/syllabus.htm
which is the Syllabus page for MBA 600. Now you need to compare
something on the Syllabus page for MBA 604. Not having anticipated that
particular need, I have not provided a direct link. It's probably four clicks
away. However, by clicking in the browser's address box next to the final 0 in
mba600 and replacing it with a 4, you can go where you need to in a couple of
keystrokes.
So how should I group the files? Again, I can use any grouping I
want. What I have done with Ricci Street is common.
As long as each of the 500+ pages on Ricci Street were in a
different directory or folder, they could all have the same file name. I expect
that you'll mix and match. While every directory needs an index.html file (or
other default page) to avoid an ugly dump page, other duplicates are up to you
as long as everything has a discrete address.
For example, you could have two homework folders, one in the
mba600 folder and another in the mba604 folder. Or you could have one homework
folder, within which you have an mba600 folder and an mba604 folder.
Tip | At Ricci Street, the directory structure
mirrors the navigation metaphor.
Ricci Street has six neighborhoods. Thus the directory structure
has six subdirectories under the main directory. Gizmos, Inc., has seven parts,
from Office to Toolkit. Guess what? The gizmos directory has seven
subdirectories.
In addition, each of the six neighborhood directories has an
images subdirectory. It contains only image (.gif, .jpg), javascript (.js), and
style sheets (.css) files, all of which are called by .htm files in other
directories. Thus, no one should be accessing any of the images directories and
none of them has an index.html page, as do all the directories
that have .htm files and people will access.
Here's what happens when you try to access one of those
directories: http://RicciStreet.net/gizmos/images/
Not pretty. Very geeky. Click around to display the images
directly in your browser. If you don't put an index.html page in every one of
your directories, this plain dump is what users will see.
If you aren't using any FrontPage goofiness like themes and shared borders,
you shouldn't need any directories that start with an underscore, such as _vti_cnf
or _private. You can probably delete them.
Delete all the empty or useless directories.
Put images and sound files in /images/ or /media/ subdirectory.
Directory names
As you can see in your browser's address window, the URL for
this page is
http://RicciStreet.net/gizmos/toolkit/webmaking/sitemanagement.htm
It has four parts
the protocol,
http://
the domain
name, RicciStreet.net, including the top-level domain .net
the directory path,
gizmos/lab/toolkit/webmaking/
the file name,
sitemanagement.htm.
The Linux operating system works best with directory names that
are all one word, all lower case. They can be two words run together, but don't
put any blank spaces. If you see the character sequence %20 in a URL, that means
the webmaster doesn't understand Linux very well and is asking for trouble.
For ease of typing, speaking, and remembering, I use directory
names that are short, pronounceable, and easily spelled. As I can see from
the server's error log, gizmos is easily misspelled gismos and
sometimes gizmoes, but at least it's short. For Parkside Plaza, I thought
using your last names was the best system. It broke down only with the two
Palmers and the two Scotts, where I resorted to adding initials.
How would you describe and characterize the structure of your
Parkside Plaza directory? Why does your directory tree look like it does? If
it's totally flat (no subdirectories), can you see any logical groupings? If the
directory tree has some depth (multiple levels of subdirectories), what's the
organizing principle? For example, the /riccistreet/ directory tree mirrors the
Ricci Street navigation metaphor. If you don't have any subdirectories yet, what
are your plans?
Files
File names
As you know, I recommend all one word and all lower case.
Example, this page is called websitemanagement.htm, probably the longest on
Ricci Street. I try to make all the files names short and meaningful. Note that
you can have more than one file by the same name as long as they are in
different directories. For example, there are lots of index.html files on Ricci
Street, but only one per directory.
In your directory /plaza/yourlastname/ count the directories,
subdirectories, and files. Many of you don't have any subdirectories. You FTP'd
all the files into the directory I made for you.
File types
A web server can be set up to deliver any file type you want. If
it isn't displayable by the browser, it will be downloaded to the user's
computer. Thus, in your Parkside Plaza web you can put and link to Word files
(.doc), Excel files (.xls), and video files (.mpg), among many others.
However, the web browsers display only .htm, .html, and other
text files such as .txt. For images, they display only .jpg and .gif. The
Microsoft browsers will also download and "display" audio files with
the extension .wav. Later versions of the Microsoft browser will display .xls
files, but you don't get any spreadsheet functions along with it. For everything
else, you need either a plug-in such as Flash for .swf files or other
appropriate software to use the files offline.
Delete all the files that aren't being used, especially the
extraneous .doc files and .mix and .wmf
Make all the file names lower case and one word. Make them all
have the form name.ext. Some of yours have name.ext.ext. No underscores or
special characters. Where possible, use naming conventions to help group files.
DCP00004.GIF is not a helpful file name. neice1.gif is helpful, especially when
it's listed right before neice2.gif. Use .html extension for index page only.
Use .htm for everything else. Have an index.html file in every directory with .htm
files.
Webkeeping
Webkeeping is like housekeeping. To be ready for guests, you
have some chores to do. They may seem like a lot right now. After you get used
to making pages, these webkeeping chores will be like second nature and you'll
do them as a matter of course.
Remember paper-and-ink reports? Webkeeping sure beats numbering
pages, printing, collating, copying, re-collating, stapling, and mailing.
Directory / folder
Be able to describe and characterize the structure of your
Parkside Plaza directory or folder. Why does your directory tree look like it
does? If it's totally flat (no subdirectories), can you see any logical
groupings? If the directory tree has some depth (multiple levels of
subdirectories), what's the organizing principle? For example, the /riccistreet/
directory tree mirrors the Ricci Street navigation metaphor. If you don't have
any subdirectories yet, what are your plans?
Put images and sound files in an /images/ or /media/
subdirectory.
Unless you're using Front Page's "extensions", special
chunks of code that I recommend you not use,
you don't need any directories beginning with an underscore, such as _vti_conf.
When you delete them from your web via FTP, remember that you have to delete the
files within a directory before you can delete the directory itself.
The file naming conventions below also apply to directory names.
Key: all one word, all lower case.
Files
Why are the files names what they are? Where possible, use
similar file names to help group files. DCP00004.GIF is not a helpful file name.
neice1.gif is helpful, especially when it's listed right before neice2.gif.
Delete all the files that you don't use, especially the
extraneous .doc files and .mix and .wmf and anything you FTP'd by mistake.
Make all the file names have the form name.ext. Beware
name.ext.ext, for example, homework.xls.htm. Is that a spreadsheet or a web
page?
Make all the file names lower case and one word. By one word, I
mean no spaces. For example, mba 504 homework 1.htm won't work on line. Windows
understands it just fine, but the server's Linux operating system and users'
browsers will choke on it. Use mba504homework1.htm, instead.
No underscores, for example, homework_1.htm. If someone links to
that page from another web page using the default link styling of purple and
underlined, the underlining will make your underscore disappear.
No special characters, for example, homework#1.htm or
homeworkfor$$@school.htm. They will choke the browser and produce error messages
for the user to puzzle over. Not good.
Use .html extension for every directory's index page only. Use .htm
for everything else.
Have an index.html file in every directory with .htm files. It
should have links to all the other files in the directory.
Learn more
NetMechanic --
"Power Tools for Your Website"
low-cost site maintenance services for small businesses on
the Web - detecting problems and repairing your Web site; online, automated,
configurable, and understandable
Netcraft's What's
that site running? - type in a domain name like riccistreet.net or ibm.com
and you'll learn a lot about the server:
what
operating system it's running on
what version of which
server software it's using
when the site was
last updated
its IP address
who the upstream
provider is, that is, who owns the block of IP addresses that includes this one?
Link frequency
LinkPopularity.com -
How many other pages link to your pages?
INeedHits.com
- How popular is your website?
page management
Web Site Management
