Underline on hyperlinks

[tangotv]

I am trying to learn HTML. I am using one of Sam's Teach Yourself books.
He says that "One of the most popular style sheet effects ...... is to remove the underlining on hyperlinks."
Is CSS the only way to get rid of the underlining on hyperlinks?

[eightfifteen]

Only other way is to use graphics for your links.

[Andilinks]

There are so many advantages to using CSS that you should begin learning it as soon as you can after you've made just one page that says "hello world."

CSS allows much more flexibility and is especially powerful if you are making more than one page with a similar theme or template.

You can then change things on all the pages at once by making a single change in a CSS file instead of changing every page one at a time which is time consuming and error prone.

Andi

[gsdickson]

CSS are awsome. I use them on even the smallest of jobs. Change one attribute of a style and it change everywhere on the site -- how cool is that!

[joeinfo]

does anyone alse have an opinion on this...

because it is so easy to hide links using css (as was mentioned in this topic), by making the link the color of the body text and removing the underline, do you think that search engines may penalize sites for using { text-decoration: none } in the styles tag?

search engines say that they penalize sites for hiding links (which are often added to pages merely to increase page rank for the "linked" sites). i've noticed this CSS abuse on many sites and -- if i've noticed it -- i'd assume that the folks at the search engines are wise to it as well.

any thoughts?

[ReviewGolf.com]

Yes, if you use textual links the only way to remove underlining is to use CSS.

If it was a good idea or practice to do, I'm sure there would've been more scripting ways.

[Andilinks]

search engines say that they penalize sites for hiding links

Search engines penalize attempts to gain visibilty through dishonest means.

Do you think they may not be clever enough to distinguish elements of style from scams?

It is possible they may not be, but unless they tell us we will never know. I don't think it's possible to test the hypothesis effectively because it wouldn't be very heavily weighted, so worrying about this is probably a waste of time.

Andi

[ronniethedodger]

Originally Posted by joeinfo

does anyone alse have an opinion on this...

because it is so easy to hide links using css (as was mentioned in this topic), by making the link the color of the body text and removing the underline, do you think that search engines may penalize sites for using { text-decoration: none } in the styles tag?


any thoughts?

Opinion? Yes.

Don't do it.

Many SE's have a "report spam" url and a lot of your competitors will use it if they spot you using deceptive practices. This goes beyond just CSS and into javascript also for the same ends of hiding or deceiving SE's for better rankings.

My thoughts? Okay.

If you are in my neck of the woods, and I see it ... I report it. I make no bones about that too.

The unfortunate thing is though, by the time that it gets noticed and the time that it gets removed is sometimes a couple of weeks to a month. That is enough time to do the damage of some loss of sales.

After the site goes down ... they pop up somewhere else and we start the whole thing over again. Sometimes Google will pick up on the practice well enough ... but then they devise another method and start the game all over again.

Eventually though, a lot of these methods will go the way of the white on white hidden text type of practice. All SE's will become savvy enough to sniff them out, that is inevitable. They can only become aware of some of these practices by individual site owners reporting them.

But to get back to the actual question that tangot raised, it all boils down to intent. If your intention is to make a more aesthetic look to your pages, then I see no problems in using text links without underlining, different colors, etc. If it is obvious enough that they are links, then go for it.

If you feel that you may get penalized for your .css stylesheets by using some of these methods, then put all of your stylesheets into a seperate directory and disallow all spiders access to that directory in your robots.txt file. The SE's will not penalize you for doing this ... it would require a competitor to report you for it (highly unlikely) and for Google or another SE to actually view your files manually to make a determination on whether or not your intent was to deceive their spiders.

[sovidiu]

Ronniethedodger is right. CSS is the only way of making the undelined hyperlinks behaviour gone and a recommended Web site building tool.

However, if you want to test the usability of your CSS, you might want to look at this Web page analyzer. I took the liberty of testing Ronnie's Web site and it returned:

CSS_SIZE - Caution. The total size of your external CSS is 2261 bytes, which is above 1160 bytes and less than 8K. For external files, try to keep them less than 1160 bytes to fit within one higher-speed TCP-IP packet (or an approximate multiple thereof).

You might want to split your CSS :)

[ronniethedodger]

Originally Posted by sovidiu

CSS_SIZE - Caution. The total size of your external CSS is 2261 bytes, which is above 1160 bytes and less than 8K. For external files, try to keep them less than 1160 bytes to fit within one higher-speed TCP-IP packet (or an approximate multiple thereof).

You might want to split your CSS :)

You are a little bugger! ;0)

I ran that test too a while back, believe me it was lot worse. But you know what, I don't care...hehehe. I am not going to quibble over that right now. I have a lot of work to do on that CSS sheet anyway .. and maybe it will get smaller. Nothing is final yet.

[Conficio]

Originally Posted by sovidiu

Ronniethedodger is right. CSS is the only way of making the undelined hyperlinks behaviour gone and a recommended Web site building tool.

However, if you want to test the usability of your CSS, you might want to look at this Web page analyzer. I took the liberty of testing Ronnie's Web site and it returned:

CSS_SIZE - Caution. The total size of your external CSS is 2261 bytes, which is above 1160 bytes and less than 8K. For external files, try to keep them less than 1160 bytes to fit within one higher-speed TCP-IP packet (or an approximate multiple thereof).

You might want to split your CSS :)

Good work sovidiu,
however, as the CSS file is reused for every page in this web-site, the performance hit is only taken once. It should therfore be of minor importance. But do the same thing with multiple graphics (rather then a common CSS or JavaScript file) and different ones on every page and you got your self into performance issues for dial -up users.

Happy surfing
K<o>

[ronniethedodger]

It does show that file transfers are done in packets and not bytes. I did not know this until a couple of months ago when it was pointed out to me.

To take the example of my poorly unseperated CSS file as an example and the 1160 byte cutoff (if that is indeed the cutoff), then a file of 1161 bytes would take twice as long to transfer as one of only 1160 bytes.

This is a good thing to keep in mind when optimizing your graphics files too.

[sovidiu]

'lo. The CSS does not need to have a maximum size of 1160 bytes; it can also be a multiple of this number (2320 or 4640, in order to be less than 8k). Anyway, a big CSS won't ruin a visitor's connection. It is also advisable to consider displaying your company's name as text with a default font (Arial, Verdana, Trebuchet, Courier, Times New Roman etc.) and just put the logo, should it represent an image not including your company's name.

Image optimization also takes in consideration the use of PNG over GIF. Most of the small images have occupy the same disk-space (e.g., a small GIF image will have the same size as a PNG one, but will have fewer colour variations). And, of course, you could cut a large image into smaller ones, to increase the site's loading time (Adobe Photoshop can make the work easier for you, the same as Macromedia Fireworks).

TangoTV, I would like to recommend you the W3C Schools, created by the World Wide Web Consortium. It's open-source and it can help you with your "iWeb" research :)

[ronniethedodger]

Originally Posted by sovidiu

'lo. The CSS does not need to have a maximum size of 1160 bytes; it can also be a multiple of this number (2320 or 4640, in order to be less than 8k).

What I was referring to is the packet size. Images, files, text, etc. are packet streamed and not bit-streamed.

You will also notice that the packet sizes are not in 1k increments also. That is why I was a little confused by the 1160 bytes boundary in the first place. I will have to look and see what they meant by that. But it seems to me that the packet includes not only the data it is sending, but handshaking information also of some sort.

[sovidiu]

I took a look at the some explanations about the way TCP/IP packets work and it proves you are right: there are three parts compounding a Transmission Control Protocol/Internet Protocol packets: the header, the payload and the trailer.

Now, the explanations in the above-mentioned Web address point that the header is three times bigger than the trailer (32x3). If you add their value (128) to the size of a kilobyte (1,024), you will get 1152 bytes, which is 8 bits shorter than the number that the optimization company suggested as appropriate (1160). So it seems that the 1160 value (32 bits header + 96 bits trailer + 1024 bits payload + 8 bits perhaps for additional information) is justified.

But anyway, the main idea is that a CSS should be smaller than 8K. You might want to compress your Web site's content using the mod_gzip compression module. It compresses the information being sent to the visitor, for faster Web site-loadtime.

[ronniethedodger]

Originally Posted by sovidiu

So it seems that the 1160 value (32 bits header + 96 bits trailer + 1024 bits payload + 8 bits perhaps for additional information) is justified.

Is that 8 bits or 8 bytes? Bytes I think.

But I do know that if you are one byte over that cutoff ... you might as well be 1000 over it. The packet is still the same size, even if there is only one byte of data inside of it.

Originally Posted by sovidiu

But anyway, the main idea is that a CSS should be smaller than 8K. You might want to compress your Web site's content using the mod_gzip compression module. It compresses the information being sent to the visitor, for faster Web site-loadtime.

Working on it. ;0)

[sovidiu]

Yes, they were bytes and not bits; you were right ;)

Here's a tutorial about Compressing Web content. And I must say that the results are very proficient (reducing the size from 12Kb to around 5Kb is a tremendous advantage). The only thing that can be regarded as an impediment is the fact that some old browsers can't handle gzipped content.

[ronniethedodger]

There is a lot of tangibles on g-zip compression. One of them is the file sizes. If they are small files to begin with, then you do not benefit from the compression all that much ... the savings does not warrant the server overhead either.

I guess there is a balance in there somewhere though. I can see if a lot of your pages are quite large to begin with (like a full page of forum posts for example) then you would benefit from g-zip compression. In fact, compression is built into the phpBB software because of this. It is an option to use it or not via the ACP for phpBB.