Poor Richard 9
#009/05-June-98
POOR RICHARD'S WEB SITE NEWS
Geek-Free, Commonsense Advice on Building a Low-Cost Web Site
Editor: Peter Kent
Top Floor Publishing
IN THIS ISSUE
- Beginner's Column: Capitalize Your Domain Name
- More Web-Site Email Recommendation Scripts
- Managing Large Link Directories
- Poor Richard's Links Page Updated
- Adding a Subject Line to a Mailto: Link
- A Macintosh Text Editor
- Poor Richard's Web Site in the Press
- Reading Back Issues
- In Future Issues
**** IF YOU FIND THIS NEWSLETTER USEFUL ... FORWARD IT TO FRIENDS AND COLLEAGUES ****
Beginner's Column: Capitalize Your URL
I've recently decided to begin writing my domain names a little differently. I used to write http://poorrichard.com/ and http://topfloor.com. Some people don't notice that there are two r's in the middle of poorrichard, though, and I think topfloor could be a little clearer, so I've begun writing http://PoorRichard.com/ and http://TopFloor.com/
I think these are easier to read. In fact you may have noticed other companies do this; they mix both uppercase and lowercase letters in their domain names. It's a simple way to make a URL easier to understand and remember, especially if you have letters repeating -- it's easy for the eye to miss two r's or two n's. For instance, there's the http://spannet.org/ site. The owners realized that the name is easier to read and recognize if they spell it this way: http://SPANnet.org/.
There's a little confusion about this, though. Many people think that domain names must be written in lowercase, or must be written in the way in which the domain was registered. So here are the basic rules.
- You can write the domain name whatever way you want, it makes no difference to a browser or Web site: http://PoorRichard.com/, http://POORRICHARD.COM/, http://poorrichard.com/, http://PoOrRiChArD.cOm/, it really doesn't matter. (Try these URLs and you'll see that they all work.)
- It _does_ matter, though, how you write the rest of the URL, the directories and filename. For instance, http://PoorRichard.com/newsltr/archive.htm will work, but http://PoorRichard.com/Newsltr/Archive.htm will not.
By the way, you can do the same with email addresses; they are not case sensitive, so if it looks better to use uppercase letters in some positions, go ahead and do so ... it won't make any difference to a mail program -- the mail will still get through.
More Web-Site Email Recommendation Scripts
In the last issue of this newsletter I mentioned a service called Recommend-It. You can put a Recommend-It button on your Web pages; clicking on the button opens a form that visitors can use to quickly send a message to friends or colleagues mentioning your Web site.
(See http://www.poorrichard.com/newsltr/008.htm#recomm for more information.)
Since then I've found a couple of similar utilities. These, however, are scripts that you install on your server. I haven't either, but they look like they're pretty simple, so if you've managed to install CGI scripts before, you should be able to work with these:
MasterRecommend:
http://www.willmaster.com/MasterRecommend/
SoftwareSolutions:
http://softwaresolutions.net/products.htm
Use the script to recommend the SoftwareSolutions Web site to someone, and you'll be shown a Web page from which you can download the script and use it for free.
Managing Large Link Directories
Many Web sites contain link directories, pages with links to other useful sites. The site associated with Poor Richard's Web Site, for instance, has a directory of almost 800 URLs that are mentioned in the book.
Now, I generally try to avoid unnecessary technical terms in this newsletter, but when discussing this subject I have no option but to employ a Web-design term. These link directories are what is known in the business as a "royal pain in the butt."
Because Web pages come and go faster than kids raiding a cookie jar, maintaining link directories is a real hassle. For instance, I recently checked some URLs associated with a book called "Career Ideas for Kids Who Love Computers" (a book I wrote a friend who owns the "Career Ideas for Kids" series). Right before the book went to print we checked the URLs ... just three months after finishing the book 10% of the URLs were bad.
Let a link directory sit for a year, and you can easily lose 20% of the links, maybe more. Yet with a very large directory checking those links by hand is a real problem. How, then, can you manage such link directories? I'll look at a few things you can do to make this task easier.
Kill the Directory
Do you really need a link directory? I remember a couple of years ago reading a review of a Web-authoring product in a computer magazine; the reviewer said "the secret to a successful Web site is creating lots of links." Complete twaddle, of course, but for a while there, in the early days of the Web, most Web sites seemed to be little more than pages with links to other Web sites. Even today many site owners throw up link directories because they think that's the thing to do.
But link directories should be avoided at all costs. If you have to have one, fine. But do you really need one? Think about whether the link directory really serves a purpose, and if it doesn't, or if the hassle of maintaining it is greater than the benefit ... kill it.
Avoid Subdirectory Links
It's a good idea to avoid links that have a good chance of dying.
For instance, if you have the choice between these two links, which is safer?
http://www.acmekites.com/
http://www.bighugeisp.com/~kites/
The first link is far more likely to endure than the second. Whoever set up the first URL went to the trouble of getting a domain name, and finding a hosting company. The second URL could change at any time ... the owner may move to another ISP, for instance, or decide it's time to get his own domain name.
A particular problem, by the way, are links to subdirectories at .edu domains. Why? Because .edu domains are colleges and universities, and most Web pages at these sites are owned by students ... who'll be leaving eventually.
Avoid Subdirectory Links Within a Domain
Even if a Web site has its own domain, you may want to consider pointing to the main page rather than down into the subdirectory structure, if at all possible. For instance, if the acmekites.com site has a clear link from the main page to the directory of kite manufacturers, you may want to link to http://www.acmekites.com/ rather than http://www.acmekites.com/info/new/manufacturers/directory.htm. That way, if the Webmaster moves things about, your link won't break.
Link Checkers Don't Work
There are many programs that will quickly check links for you. In some cases these programs are even built into Web-authoring programs. For instance, FrontPage has a utility that checks all tech links at your site to see if any are broken. Now, this is useful for checking internal links, and a useful quick check on links, but it's really not a good way to check a link directory.
These link checkers simply try to obtain a page from the server, using the information provided by the link; the domain name, directory path, and filename. If they get something, that's fine, the link is okay.
But what if one of your links currently points to a page that says, "We've moved, please change your bookmarks to this new URL"? The link checker tells you that the link is okay. Or what if the link points to a page that is completely different from the one you originally linked to ... that is, all the relevant information has gone? Again, the link checker tells you it's okay.
You can't rely on link checkers. Instead, you need an off-line browser.
Use an Off-line Browser
An off-line browser is a program that will grab Web pages for you automatically. You simply point it at your link-directory Web page, and let it run (overnight, perhaps, as it may take several hours). The program follows all the links on the specified page, down as many levels as you tell it to go. When it's finished you can view the pages very quickly, because they're stored on your hard disk. That is, you can view the pages "off-line."
For the purposes of checking a link directory you only need to go down one level; in other words, tech off-line browser will grab the page pointed to by each link in the directory. In a future newsletter I'll discuss doing research with an off-line browser. When using it for research you may want to go two or three levels down; that is, the program grabs all the pages pointed to by the first page (one level), all the pages pointed to by _those_ pages (two levels), and all the pages pointed to by _those_ pages (three levels). But going down multiple levels can be very slow, and use up huge amounts of disk space. For checking a directory, one level is enough.
I use a program called Netattache. I looked at off-line browsers some time ago, and thought most were awful. They were poorly designed, and in some cases had glaring faults. Flashsite, for instance, was actually quite well designed. But it had the most amazing omission; it would only grab Web pages that were stored at the same domain as the top-level page. So Flashsite would be completely unusable for checking a link directory. (This is a classic case of programmers asleep at the keyboard; not only does it seem obvious to me that a program must be able to grab pages from other Web sites, it was obvious to all the other companies creating off-line browsers.)
Whatever off-line browser you choose, it really needs these features:
- It must be able to dial into your service provider automatically if the line drops; it may take hours to check your link directory (it took about 5 hours to check my directory), so if the line drops the program must be able to dial back in and continue.
- You should be able to define what sort of data is transferred; I tell Netattache not to transfer Java apps and multimedia, for instance, though I do transfer pictures.
Netattache has a very useful feature; tech ability to look for changed pages. That is, it can check your directory completely, then, the next time it checks, it can transfer just the pages that have changed.
If you want to try Netattache (you can use it for free for 15 days, then it's $49.95 to register), visit this Web site:
Here's how I use Netattache. I point it at my link directory, and tell it to go one level internally, and 1 level externally. Internally refers to pages stored at the same Web site as the specified page; actually I don't really need to go any levels internally, but Netattache won't accept 0 for this setting. (I think there's a bug somewhere here, though; sometimes Netattache just won't grab pages externally, and I've found that setting the _internal_ levels to 2 will force it to do so. Oh, and another bug to watch out for; don't install Netattache below the /Program Files/ directory; rather, install it in the directory it suggests -- if there's a space anywhere in the path, Netattache won't start the Netattache server, so it won't be able to display the information it has collected from the Web.)
Then I let Netattache run overnight. In the morning, I open up the "brief" (that's the term Netattache uses for the collection of Web pages it gathers). The brief opens in a browser window, and shows me my list page. I then open another browser window, and size the brief window so I can move it to one side of my screen. I then drag the first link from the brief window onto the other window.
Did you know you can drag links from one browser window to another, at least in Windows browsers? You can do so with Netscape Navigator and Internet Explorer; it's a great way for opening multiple pages from the same list, as it's much quicker than using the Back button to return to the list each time you're ready to open the next page. Point at a link, then click and hold the mouse button down, then drag the link over the other window and release the mouse button. Only one problem ... if you drop the link into a frame, the document loads into that frame, so you may have to remove a framed document first by loading a blank document -- I created a button in my browser's toolbar that loads a blank page so I can quickly clear framed documents.
As the Web pages are on your hard disk, they should load very quickly. Now and then, however, you'll see a message saying that Netattache couldn't grab the page; here's a message from my last check:
---------------------------------------------------------------
Inaccessible Page
Netattache Pro was unable to load www.cocacola.com/. Either the site does not exist or the network is too heavily loaded at this time and Netattache Pro could not access the page.
---------------------------------------------------------------
As the message points out, the page might be there. In fact in this example there's definitely something at www.cocacola.com. Sometimes Netattache can't transfer pages for other reasons; it doesn't like cookies and redirect pages, for instance, and doesn't work with links pointing to FTP sites or with mailto: links (links that open an email program).
For instance, the page pointed to by www.cocacola.com redirects tech browser to another page, and this seems to confuse Netattache. (The Coca Cola page uses this tag: <META HTTP-EQUIV="Refresh" CONTENT="0; URL=/cokecard/">. I discussed using Refresh in issue 5 of this newsletter; see http://www.poorrichard.com/newsltr/archive.htm ).
If Netattache is unable to grab a page, just click on the link (www.cocacola.com/ in the Inaccessible Page message above is a link) and see what happens. For some reason you'll occasionally get a blank page; no Web page, no error message. You'll just have to check these in your link directory manually.
When you gather information about your links, you'll find that a disconcerting number need changing or perhaps even removing. You can modify your link-directory Web page directly, or use a Bookmark program.
Working With a Bookmark Program
I'm using a bookmark program to store all my links. A bookmark program is one that will store browser bookmarks -- that is, you can use it instead of the bookmark or "favorites" utility inside your browser. Ideally the program must be able to do these things:
- Load Web pages in the browser you use.
- Load Web pages in both the popular browsers: Netscape Navigator and Internet Explorer.
- Create Web pages from the bookmarks.
- Create Netscape Communicator bookmark files from the bookmarks.
- Create Internet Explorer Favorites folders from the bookmarks.
I'm currently using a Windows program called QuikLink, because it's the only one I've found that can do all of the above. However, there's a problem with this program; there's a bug that messes up program switching, at least in Windows NT (the operating system I use). When QuikLinks is running, switching between programs is awkward -- QuikLinks keeps jumping to the front, and makes other programs flicker badly. Sometimes it even seems to temporarily lock up other programs, so that when you switch to them the buttons don't work for a few seconds. I don't use Alt+Tab at all to switch between programs when I'm working with QuikLinks -- I just click on the program button on the taskbar to switch to a program, and that helps the situation a little.
All this is very irritating, but it's the only program I've found that can do what I need. And perhaps it doesn't have this problem in Windows 95 or 98 (I haven't checked). I'm still hoping they'll fix the bug. Apart from this bug, it's a great program. It looks a bit like Windows Explorer -- it has a directory structure with folders and subfolders. And each "file" stored in a folder is a bookmark -- each bookmark represents a link in my link directory. So I can quickly change the information for a link by modifying the information in a bookmark.
When I've finished, I can use the QuikLink WebWizard to create a Web page, which I then post at my Web site. If you wish you can also use the program to create a Navigator bookmarks file, or a collection of Internet Explorer Favorites folders, so people can download them from your site (I haven't chosen to do this, though I'm considering it).
For more information about QuikLink, see ...
The program is free for 30 days, then $40 to register. Remember to download both QuikLink Explorer Gold (you need the Gold version in order to create Web pages), and the WebWizard add-on (which is free). There's also another add-on, AutoBot, which can check Web sites, and inform you when one of the bookmarked pages has changed.
Removing Pages
You'll often find links that have changed; they're easy to deal with, you simply replace the old URL with the new. The problems are the links that you can't connect to at all; are they still there? Perhaps the server is down for a little while, or even a few days. When I run into this situation I quickly check with Whois. In fact I have a button on my browser's toolbar that takes me here:
http://rs.internic.net/cgi-bin/whois
This is the Whois service at InterNIC, the organization that handles .com, .net, and .org domain names. I type in the domain name and press Enter. If the domain name is no longer listed, the link is truly dead. If the domain name is still listed, Whois displays the contact information for that domain. Now, if you really must have a perfect directory, you could contact the organization that owns the domain and see what's going on. Otherwise you could give the link the benefit of the doubt and assume it will be working eventually ... and leave it in the list. Or perhaps come back and check later.
By the way, if anyone knows of good Macintosh off-line browsers and bookmark programs, let me know and I'll mention it in the next issue.
Poor Richard's Links Page Updated
Well, after the previous section it probably won't surprise you to hear that the Poor Richard's Links Page has been updated. This is the page that contains all the links in the book, "Poor Richard's Web Site: Geek-Free, Commonsense Advice on Building a Low-Cost Web Site."
Lots of links had changed, and in most cases I've been able to modify the link to point to the correct information. In some cases the link is simply dead ... but there again, I've also added a few new links. You can get to the page here:
http://www.poorrichard.com/links/
This is a very large page, by the way; it contains hundreds of links (almost 800, links to services useful to Web-site owners and administrators, from places to learn about Web design to companies that provide credit-cart merchant accounts). I'd recommend that you save it on your hard disk, and use it from there.
Adding a Subject Line to a Mailto: Link
You probably know how to create mailto: links. These are Web-page links that open an email program when the visitor clicks on them. Just in case, here's one:
<a href="mailto:subscribe-prwebnews@lists.lyris.net">Click here to subscribe</a>
If a visitor clicks on this link, his mail program opens, with subscribe-prwebnews@lists.lyris.net in the To: box of the message-composition window. Now, this is the address that is used to subscribe to Poor Richard's Web Site News. I'm using a system called Lyris, and with that system you can send a blank message, with nothing in the Subject line, in order to subscribe.
But many systems require the word SUBSCRIBE in the subject line. Well, you can make the mailto: link place the word SUBSCRIBE in the Subject line. Simply add ?subject=SUBSCRIBE to the end of the mail address, like this:
<a href="mailto:subscribe-prwebnews@lists.lyris.net?subject=SUBSCRIBE">Click here to subscribe</a>You can replace the word SUBSCRIBE with any text you want, of course, but this is a handy way for creating a subscription link for a newsletter or mailing list.
A Macintosh Text Editor
In the last issue I mentioned TextPad, a great little Windows text editor that I'm now using to write this newsletter ... as a way to avoid the problems caused by non-text characters, inserted by a word processor, ending up in my messages.
Several readers told me that the text editor to use on the Mac is BBEdit. There are two versions, one's free, and the other's $70.
For more information ...
Poor Richard's Web Site in the Press
With journalists requesting around 600 copies of "Poor Richard's Web Site: Geek-Free, Commonsense Advice on Building a Low-Cost Web Site," reviews turn up all the time. Here are a few we've received recently ...
- Jennifer Buckendorff, reviewing Poor Richard's Web Site at Amazon.com, evidently likes the book. "With all great new things comes a proliferation of hucksters and snake-oil salesmen, and the Internet is no exception" she said. "The antidote to this swirl of confusion lies in Peter Kent's 'Poor Richard's Web Site.'"
- Watch for "Poor Richard's Web Site" on the Disk Doctors TV show; it's going to be featured sometime in the next few weeks.
- Poor Richard will be reviewed in The Library Journal any day now, and should be in Technical Communication (the journal of the Society for Technical Communication) and NACE Journal soon, and recently appeared in MicroTimes, too.
For more reviews, and many readers' testimonials, visit ...
Poor Richard's Web Site is in many bookstores, and can be ordered by others; it can also be ordered on-line, through the mail, by fax, or by phone.
See ...
http://www.PoorRichard.com/order/
Order direct from the publisher, and you'll get a 100%, 1-Year Guarantee. If the book doesn't help you set up a low-cost yet effective Web site, send it back for a refund!
Reading Back Issues
If you need to refer to back issues of this newsletter, you can find them at the following location:
http://www.PoorRichard.com/newsltr/
In Future Issues ...
- Getting the Word Out about your Web site
- Setting up an email mailing list for newsletters, bulletins, product announcements, etc.
- Secure servers: why you want one, how they work, who has them
- Mail-merge programs and mailing list programs
- A quick way to inform Web sites when you change your URL
- Registering your newsletter with newsletter directories
- Are people cheating at the award sites?
- Charging purchases to the buyer's telephone numbers
- Electronic press releases _do_ work!
