Transferring Files to the FTP Server

Copying Data Onto the Server
If a network with a direct connection to the Internet is not available, HTML documents will need to be created and tested on a local computer. Once finished, the HTML files must (diagrams etc.) be transferred to the Internet server. In order to complete this step, a dial up connection to the Internet is needed. However, dialing directly into our server is not necessary.

The communication with the server is accomplished by utilizing a special type of software called an FTP client. The primary function of such a program is to transfer files from a local computer to the Internet server via FTP (File Transfer Protocol).

There are many shareware providers on the Internet who distribute different FTP programs. The most well known FTP client is WS_FTP (http://www.ipswitch.com). Using WS_FTP as a model, we will describe a standard FTP session.

Working with WS_FTP

As with Internet browsers, FTP clients assume that there is an active TCP/IP connection. If this connection is already established, it can be changed through the task manager in WS_FTP. This means, for example, that browsers and FTP clients can be active at the same time. If there is no TCP/IP connection to the Internet, it is normally structured as soon as WS_FTP is accessed. Whether or not the automatic selection will function depends on your system configuration.

Once WS_FTP is open, the window "session profile " will be displayed. Here the access information for the FTP connection is entered.

Push the "New" button and then enter the following information:

Profile name: A term selected by the user that describes a particular connection to the server, e.g. domain name or IP address.

Host name: The domain name. If the domain name is not registered, the account IP address can be entered into this field.

Example:

Host Name: 207.201.156.106

Host type: The servers at Road Runner, Inc. run on either UNIX or NT platforms, therefore the selection should be "Automatic detect".

User ID: The user ID of the hosting account should be entered into this field.

Password: This field requires the same password used to access the account control panel. If the small box titled "save passwords" is checked on the right, the password, along with the rest of the session profile data, is going to be stored for future use.

Account: This field can remain empty.

Remote host: A directory name on the server can be specified here. WS_FTP enters this directory immediately after the Login. By default it is left blank, which implies the root directory.

Local PC: A directory name of the local PC can be specified here. WS_FTP enters this directory immediately after the Login.

Push the Save button to store the entries.

Push the Ok button to establish the connection with the server.

After the login, the main window of WS_FTP appears. The window is divided into two halves. In the left frame, the directory on the local disk drive or local network is displayed. In the right frame, the current directory on the Internet server is displayed.

In both directories navigation is performed as usual. Directories can be changed using (ChgDir) or created using (MkDir). Files can be deleted with (Delete). Text files can be loaded immediately into an editor (View) where they can be worked on and saved locally. Files can be executed using (Exec).

The (Refresh) and (DirInfo) buttons are particularly useful when working with a remote server. It can be beneficial, with an extremely slow data connection, to turn off the automatic refresh option. Otherwise, each individual modification will require the update of all the directory data. If the automatic updating of the remote directories is switched off (option/session option), the directory can be manually refreshed using Refresh. (DirInfo) will display the file list in more detail, showing the specifications of the user rights.

The most important buttons are the two arrow buttons in the center of the WS_FTP window. With these arrows, files can be transferred from your local computer to the remote host and vice versa.

It is important that the file transfer takes place in the correct mode. The transfer mode is adjusted to the appropriate option from the menu located under the directory windows:

ASCII: ASCII is used for HTML files, text files and CGI scripts.
BIN: Binary mode is used for programs, pictures, text processing files etc.
Auto: Using the Auto feature will permit WS_FTP to detect and use the proper transfer setting for each particular file. This must be configured under "options".

Error Detection Tips

After copying the data onto the server, the modifications should be tested immediately. Occasionally, pages that functioned locally do not work on the Internet. Included below is a list of the most common scenarios that can prevent a file from being displayed on the Internet.

The "index.htm" File

If the site is accessed through the browser without the explicit specification of the file name or the sub directory of the domain, the index of the directories will be displayed.

The display of the table of contents of the site will list all files and directories. Visitors to the site may have access to certain files that are not intended for the public. However, if an HTML file is created in the root directory with the name index.html or index.htm, the browser will automatically display this page. In this case it is no longer possible to get the directory listing. Therefore, it is recommended that an index.html file be stored in each directory.

Tip: Store an index.html file in each directory. The homepage of the Site should be stored in the root directory and be named index.html. This will cause the automatic loading of the homepage, domain.com/index.html, even if the visitor enters only your domain name as a URL.

Note: With each newly created domain account a standard file named index.html is installed into the root directory. Even with the presence of index.htm the .html version has priority. If the homepage is named index.htm, the pre-installed file, index.html, should be deleted.

Case Sensitivity

Conflicts in case usages are common errors when requesting files on Unix systems. The problem occurs particularly in the use of Windows 95, where the file designations differentiate between upper and lower case, but applications ignore case when opening files.

In addition to this, the user of Windows 95 does not have complete control of the file names. For example, in DOS, file names in the Explorer begin with an upper case letter. Many HTML editors create their links on this base. WS_FTP takes over the original DOS convention of writing, the result being that the appropriate links are not executed on the server. Thus, it is important to pay close attention to the way file names are written.

Tip: Always write file names in lower case characters.

Absolute and Relative Links

A common cause of broken links is the use of incorrect file paths. This scenario can arise if the HTML editor used to design the site calls files using exact paths to their location on a drive. During local testing the site will function, but, upon publishing to the server, the links will be referencing the wrong location. Some editors enter an absolute file path, which is a path on your local computer.

          file:///d:/myactualsite/test/file.htm

If a file with such a reference in a SRC or HREF attribute is sent to the server, the browser actually finds the requested HTML page, because the file is actually there. However, it is only on the computer on which the site was developed. Calling the file from another system will result in broken links. To ensure that this does not occur, SRC or HREF attributes should be coded as follows:

          <a href="http://domain.com/test/file.html">
          <a href="test/file.html">

The cache is a local storage area in which loaded HTML pages and diagrams are stored. If a loaded page is accessed again, then the copy is loaded from the cache. With this technique unnecessary Internet traffic is avoided. However, cache must be refreshed once there are new versions of cached files on the server. To ensure that the current copy is loaded from the server and not from the cache, delete the disk cache and the memory cache.

Tip: If the Shift button and the Refresh option are pressed at the same time, the page will be requested from the server and not from the cache.

This caching issue can occur in proxy environments. The Expires feature in the HTTP-EQUIV attribute of the META tags can be used to define an expiration date, after which the page is loaded again from the Internet server.

<META HTTP-EQUIV="Expires" CONTENT="28FEB1999">

This tag is intended for information such as stock exchange queries. It should be used with consideration because switching off the cache results in higher loading times for visitors of the site.

Compressing Files For Transfer

When using a relatively slow Internet connection or transferring large quantities of files, it is recommended that files be compressed or ZIPPED. Compressed files can be transferred to the server and then decompressed for use. The files can be compressed with PKZip or with WINZip (www.winzip.com).

When unzipping files certain specifications must be entered into the HTML form, which can be accessed through the browser at the following address:

http://www.domainname.com/cgi-bin/secure/archgw

Replace www.domainname.com with your own domain.

Input filename: Indicate the complete path of the file, starting from root and including subdirectories.

Output directory: Indicate the names of the directory into which the file will be unzipped. (e.g. test /).

Archive Options: Files can be compressed in different ways:

We assume files are compressed in the ZIP format. In this case select:

DOS/Windows pkZIP [.ZIP]

Compression Options: If data was compressed with PKZIP or WinZip, select the following options:

Compression Options: None

Now, files can be decompressed using " extract now ".

Important: The file names are unpacked on the server in the same way that they are in the archive - this means with the same upper and lower case writing. WinZip is recommended because it allows the checking of spelling for the file name.

If there are path names stored in the archive, the appropriate directories are created during decompression. An entire site, with all directories intact, can be extracted to the server using this feature.

Announcing Sites to Search Engines

Once the site has been transferred to the server, it is ready for viewing. In order to generate site traffic, the Internet must be made aware of a site's existance.

If someone needs information on the Internet, they often use the popular search engines. Essentially, there are two types of search engines. The first type is the Web-Spider (robots), which automatically go from link to link in a site and analyze information. The most well-known search engine of this type is Alta Vista (http://altavista.com). The second type indexes the content of registered sites. The most well-known search engine of this type is Yahoo (http://yahoo.com). Additionally, there are combinations of the two search engine types.

To be listed on a search engine of either type, a site must first be registered with that engine. Since there are many search engines, registration can be a time-consuming affair. Services such as Submit-It will submit a site to the most important search engines (http://submit it.com).

With some search engines a META tag can influence the indexing of the site. These tags usually contain a brif desrciption or a list of key words.

How the information in the META tags is analyzed depends strongly on the search engine.

Normally, it is desired that all pages are reachable using search engines. However, occasionally it is necessary to prevent certain information from being included. A development directory used for testing is one instance where content should not be listed. The exclusion of directories is achieved through the creation of a file named robot.txt. In this file, directories that are not to be processed by search engines are listed.

The robot.txt file (the file name must be in lower case) must be in the root directory of the server. It has the following structure:

After "Disallow:" the directories which are to be excluded are entered. Under "User-agent" special search machines can be indicated. Those indicated will have the restriction applied. (The star denotes all search machines). Several such definition blocks can be indicated, separated by a space. The access of individual search engines can be controlled in this way. Another possibility is the META tag.

The INDEX directive indicates, whether or not a robot is to indicate a page. The FOLLOW directive determines whether or not the search engine should follow the link on the page. The defaults are: INDEX and FOLLOW.