Demographics of Web Users
Submitted by atcooper on October 28, 2003 - 21:25
When doing research for my own webcomic I found this page.
Take note of the real download time of those who use dial-up. That's 40k. The 56k is an ideal connection speed.
I was not sure wether this should go under buisness or theory. I chose theory since I was thinking more in terms of how to design my pages when I was looking for this information.
- Array
Some elaboration
by atcooper - 10/30/2003 - 14:08
I realize I did not spell out my conclusions from this.
One, I think the potential audience for comics on the web could be expanded if more authors would try keeping the file sizes smaller. I too love the potential of color without the expensive printing, but massive amounts of colors still bloats loading times.
I myself need to do a little more research on file compression, but if I understand justinpie, you can get small sized files with gif if you keep the colors under 256. This needs to be considered from the start when drawing up your pages. A lot already do, but many do not.
Two, if you break the words out of image files and use html to process the text, then you can take advantage of all the other readers globally who do not know english. Many of these translation engines are not yet powerful enough to give really good translations, but you can get much bigger audiences this way.
This will require positioning code, which requires a little more familiarity with hand coding. Ridding yourself of the bubble convention makes things easier, too. I am not yet familier with the many wyswyg programs out there, but I see in my mozilla composer that it can't be done without actually going into hand coding.
I know a great majority of people into web comics have dsl lines. As more time goes on, more people will sign on with dsl, but if comic fans are equally distributed between dial up and dsl, then anyone who designs with only dsl in mind is shutting out sixty percent of the users on the net.
I know writers who would mock me with the translation mangaling thier carefully crafted poems, but most comics are not word sophisticated enough to worry about such issues. Humor is harder to translate as well, so many jokes would be lost on non-english speakers. Hemingway, who has a straight forward style, is easier to translate than Faulkner, who is anything but straightforward.
Please do respond if objections come to mind.
by David Wright - 10/30/2003 - 15:04
Interesting ideas. I don't know if breaking text out of the comics would save load time, as the image area would probably not be much smaller, as text is usually done over background which is somewhat necessary in a comic. Add to that the html and whatever other coding you would need to do to get the font just right, positioned correctly, etc....and it might ADD (though a tiny bit) to the size of the page.
As for the file sizes, those are good points. It's tough to find the perfect balance between quality of the images you are doing versus the accessibility of the slowest internet connections.
http://www.toddandpenguin.comhttp://www.takingupspacecomic.com
by Anonymous - 11/01/2003 - 23:28
toddandpenguin,
Css-p allows text over image. It's a little strange how it's done- I spent a good portion of a day trying to make it work. If you are curious as to how css works, and how it can actually reduce file size, I would be happy to give a brief overview. I have noticed a lot of people believe that css can make loading time longer, but it doesn't, as long as it's used right.
A lot depends on the stylizing of the text. There a lot of things that can't be done with html yet. I have no doubt this will change a lot in the upcoming years, so, for now, anyone using strict photoshop and/or flash will be able to do more.
The trade off is in abandoning all the collective power of applications already on the net. I cannot name all of the translation engines available on the net. It was Google's translation service for english speakers that got me thinking about this issue. It utilizes the straight html text, and cannot access words that are actually image files.
Really it all depends on how much of the net's resources you want to be able to utilize.
I see this like I see the transition from strict acoustic guitar to electric guitar. It is very well to make the newer sound like the older. The sound didn't become innovative until people began treating the electric like its own instrument.
I see the net the same way in the print versus net dicotomy.
In truth, I am working on something of my own so that I have something to point to, something to show to illustrate my points. Hopefully I will have something in a few weeks.
I appreciate you responding, by the way. It made this board feel a little less distant.
by atcooper - 11/01/2003 - 23:34
That was me on that last repsonse. I will make sure I am logged in so that I am not identified as a guest.
by eldritchmonkey - 11/02/2003 - 15:30
The only problem I could see with that would be making use of regional dialects in dialogue, IE deliberately mangeling the English language. Everything would have to be grammaticly prestine and spellcheckable. (Spellcheckable should be a word!) Who actually talks like that?
by atcooper - 11/04/2003 - 19:28
Yeah, dialects will get mangled. That's some of what I meant about Faulkner being harder to translate than Hemmingway. Simple sentence structure with common words will translate easily. Slang and idiomatic language won't though. And look out if you play with grammatical conventions.
by TCampbell - 11/05/2003 - 14:57
Also, webcomics have great dialogue, but a search engine can't read them. With Atcooper's suggestion, it could. My comics tend to play with casual speech a lot, so I'm more interested in searchability than translation.
Also, file size depends on more than the physical size of the image. Text that is rendered as an image usually has a large file size, because of all the twists and turns you need to make recognizable letterforms. A comic with HTML dialogue would almost certainly load faster.
by atcooper - 11/08/2003 - 01:26
TCampbell,
That was one of the spurs that got me thinking about breaking the words out of images. The search engines can't access the words that are actually images. I went with the translation issue primarily because I didn't know enough about how search engines work, (and they all kinda work differently), and because I know a lot of comics get read in the East.
Be wary, though, about just how much time can be wasted trying to get your site high up on any particular search engine. My impression is that you have to have a distinctive name for the site, and have lots of people linking your work. No doubt there are other factors as well. (Like people knowing about the comic in the first place.)
The main thing, though- the search engines aren't going to see the words you use in the comic if they are embedded in the image.
Css is a lot easier to use than people realize. You get a whole lot of control over how the broswer renders a page, too, without having to resort to making the whole comic one set of strung together image files.
by Anonymous - 11/09/2003 - 09:30
The problem with breaking words out of images ultimately comes down to positioning.
Unless you decide to use CSS to place everything *exactly* in one area on a page, probably using point sizes in order to have everything line up perfectly with the images in your panels, getting the text to actually stay where you want it is going to be a pain.
Combine this with other text placement problems -- what if the person has their default browser setting set to *override* css text sizes, thereby making all your placed text run into each other, and run over your images, making it a big jumbled mess? I know when I'm browsing in Mozilla I prefer to have my own font set to override whatever the designer has chosen for his/her site.
Using CSS to positiong your text over images is possible, even barring those setbacks, but it will also take a bit of trial and error to position them in the right place, unless tools like Dreamweaver actually let you do it by sight. But Dreamweaver is a fairly expensive piece of software these days...
I'm all for using Cascading Stylesheets -- it helps make a cleaner website and significantly cuts down on bandwidth -- but for somethings the extra time involved doesn't provide the payoff. Using CSS for text is one of those things that would add an extra hour at least to each strip, and I'd rather not.
Chris Wright (wrightc@ubersoft.net)
Help Desk (http://ubersoft.net)
by atcooper - 11/12/2003 - 01:24
An extra hour for the payoff in search engines alone seems worth it.
Most drawing programs will give pixel location, and from that point on it's simple math. Anyone who can balance a checkbook will be able to do it.
The overide is a more difficult issue. Designing for the web requires a certain flexability to account for user overrides. This looseness of design is very anti-print. I figure that is why most are using images only- and why flash is so popular as well. The design has to account for the variation in font size. There is a definite range, though, so it's not as amorphous as it seems.
by Anonymous - 11/12/2003 - 11:53
You say it seems worth it -- I say it seems a pain in the ass :)
I used to use three different programs to create help desk -- one was used for placing the pictures, one was used for placing the text, one was used to clean up the finished result.
other than the amount of time it took, the main problem was trying to guess how much space in the comic the words were going to take up. It often became apparent that I had underestimated how much space I'd needed, and I had to rewrite the dialog to make it -- just barely -- fit. And so the text usually looked crowded. If I wanted to fix that I usually had to go back and re-draw the damn comic.
At any rate, sure, there benefits to doing that, but I still don't see the benefits outweighing the inconvenience...
Christopher B. Wright (wrightc@ubersoft.net)
Help Desk (http://ubersoft.net)
by atcooper - 11/15/2003 - 13:30
Oh, convenience isn't everything.
I try not to be over-perfectionist, but using straight image files goes against the grain of everything else going on in the web.
by Steve_Hogan - 11/18/2003 - 10:40
I appreciate that atcooper is giving us food for thought. I do have some counter arguments.
With regards to graphics it's something of a toss up and depends on your ambitions. One of the advantages of webcomics is not being constrained by the limitations of printing expenses. Strips that can do things that are hard for print comics may have more of a "Wow" factor. Consider the fact that the most succesful video games aren't the ones that play on the most systems, but instead are the ones that look the coolest.
While an intriguing idea,doing text as a seperate layer of true text may alienate readers too. I hear a lot from traditional comic book readers who aren't crazy about computer fonts in comics and especially fonts that don't look natural in the context of the strip. Until the point is reached where people can view any font on any computer without having it installed on their computers, this would be an issue.
I question how many people who can't read english will go through the trouble of running your text through a translator. My experience with my own strip is that outside of most English speaking countries, I tend to skew towards Western European countries that traditionally read a lot of comics. I'm not sure if that would be radically different with this system. For instance, I've gotten far less hits in Ireland than France or Belgium...two non english countries that read a lot of comics. (Even though Alan Moore lives there now, there really isn't a strong comic book tradition in Ireland.) Most Western Europeans can read english any way. That leaves Asia, which is a tough market to crack and who's languages are harder to translate. In most of the rest of the world, comics are not a big deal.
With regards to search engines: I can appreciate the value of that kind of traffic to a point (I get a lot of people looking for kegs)...but wouldn't it just be easier to include all the text in your meta tags?
[url=http://www.acidkeg.com/][img]http://www.acidkeg.com/akbanner.gif[/img][/url]
by atcooper - 11/19/2003 - 15:06
Steve_Hogan,
The Wow Factor: This I can appreciate, up to a point. I have not read Megatokyo nor all of Nowhere Girl precisely because it takes too long to load the pages. Something that would take ten minutes to read in print takes closer to thirty or maybe even an hour on dial-up. The main advantage I find to internet comics is distribution. I am something of a minimalist, though- from taste, and from the more practical limitations of file compression limits.
Seperation of text from the image: I actually got to thinking about this because of my freinds in the IT and programming industries. I can see it alientating readers of traditional comics because of the changes to conventions. There is no easy answer to this. For me, I hope the audience who likes experiemental work is large enough.
Translation: For me, it's taking advantage of the automated systems already in place. Getting the word out is going to be the bigger challenge at that point.
Search engines: Again, getting the word out is going to make all the difference. I honestly don't have a clue what difference being search engine capable will do. Including all the text in meta tag's, though, will slow loading time, and some engines might not even look at it.
With the research I have done since I posted the first entry I think the loading times issue that all this was prompted by comes down to two things: image file size and page construction.
Image file size is an easy one. People who don't mind cutting out the dial-up's don't have to worry about this as much. For those who do want the dial-up's included, though- they have to try to keep file size as low as possible.
Page construction is more difficult. The more you know about how html and css and the various scripting languages work, the better. There is a whole lot to know, and I myself have only really begun. I cannot stess enough the advantage of css over tables for layouts. Take a look at brainjar for some accessible info on css. Even without positioning for removing the text from the images, it could be very useful in laying out the page with trimmer (fast-loading) code.
And after writing and replying and reconsidering, I am really glad I posted this under theory. Until I have something posted and see how these ideas actually work out, it all ends up being just theory.
by Al Schroeder - 11/19/2003 - 17:06
[quote:8c30b3ff5d="TCampbell"]Also, webcomics have great dialogue, but a search engine can't read them. With Atcooper's suggestion, it could. My comics tend to play with casual speech a lot, so I'm more interested in searchability than translation.
Also, file size depends on more than the physical size of the image. Text that is rendered as an image usually has a large file size, because of all the twists and turns you need to make recognizable letterforms. A comic with HTML dialogue would almost certainly load faster.
Ye gads.
No need to reinvent the wheel, guys.
Nothing to it. All you need to do is put the dialogue in alt tags with images. If you'll notice on my site if you hover the mouse over the image the dialogue appears. I do it all the time....and show up in the STRANGEST search strings.
In the HTML, after the image information, put alt="Mindmistress: What do you mean I can't treat President Bush as a supervillain? Why NOT? Oh. Because Vice-President Cheney is the real villain. I SEE..." and close it, and there you have it---ALL your dialogue is trackable in Google or other search engines.
Basic HTML.
Do a search for MINDMISTRESS in Google. The second entry has dialogue..."Oy vey, I wish I had hiking boots"---then follow the link. It takes you to the page where the dialogue ONLY appears in the image---unless you put your mouse over it.
I come up in all SORTS of searches. I love it!----Al
Al Schroeder III of MINDMISTRESS---think the superhero genre is mined out? Think there are no new superhero ideas? Think again.
Someone has done it
by atcooper - 11/20/2003 - 20:57
I found this via Mr. McCloud.
Bohemiandrive has pulled off some of the things I had hoped for. Here is a comic that uses positioning, and made it scalable. They still keep the words as image files, but it might be more desirable to do so- default fonts sometimes clash badly with the illustration style.
It is very enheartening to know some of what I have in mind is possible. They have hidden away a lot of the pertinant code, like the specific css rules they use, and the scripts they use, but I know (sorta) how they did it.
To see it in action, change the font size. You'll see the panel change size proportionally to the text height.
Very exciting stuff.
by Alec - 11/21/2003 - 00:47
atcooper,
Hey, that's us! :)
(Got your message by the way.)
Though my brother Adam is the writer/artist, I'm the technical guy, in charge of all the html etc.
I basically (very basically) wrote a php script that remembers, using a database, where the panels are supposed to go and how big they are in pixels and then converts all the pixels to ems (at the rate of 16 pixels to one em with two decimal places) and generates the html (<img> tags inside <div> tags, the <div>s for positioning; it's possible to do it with just <img> tags but I was worried about browser peculiarities). Once you've set the positions and sizes in ems, the browser does all the work.
It might even be possible to do the conversion with a javascript ...
About the captions, initially I was adamant we use actual text instead of images (for the benefit of search engines and because of the considerable savings in bandwidth, not to mention the ease of editing), but it turned out that it's very hard (probably impossible) to control the look of it if you're allowing the text size to change--especially, in my experience, in Internet Explorer. For instance, if you define the caption box to be 10em wide, and you make sure the text fits neatly on two lines, you'll find that if you dial the text size up or down a little bit the text will end up expanding to three lines, unpredictably, or you'll end up with a really wide gutter. If you broke the lines manually (with a br) it might be doable, though imperfect, but it would take a lot of testing for every caption (and on multiple browsers).
We don't use the alt attribute of the img tags to embed the text in search engine-friendly form basically because of the tedium of typing it all in. :P
Most of the javascript has to do with keeping the navigation bar centered on the page and sized correctly while preventing it from spilling off when you change the text size or the size of your browser window (it was pretty tricky). I would have saved myself a lot of headache if I had just fixed the size of the navigation bar (defined the size in pixels instead of ems, basically) and/or used tables for layout, but I can get kind of obsessive about programming and such. I tend to take the hard way out. :)
by atcooper - 11/22/2003 - 20:02
Alec,
I gave a lot of thinking to how you did what you did with the page. I think it's wonderful that your brother had the sense to get your help on this. I have been teaching myself the code for the last couple of weeks. Scripting seems to be where all the finesse comes in, and where all the biggest headaches are too, especially with cross-browser consideration.
I appreciate you spilling the beans on your linked files. I may use some variation of the scaling panel myself. I want to try making the design flexable enough to accomidate the problems you mention with pure text. My experiments are mostly failures at this point, but I've been questioning conventions as I've gone along, and the only thing I fear now is that if I get it to work, it may not resemble comics enough.
Positioning requires some weird nesting like the div solution you mention. That's what made it so hard to understand how positioning actually worked.
On Words Getting the Shaft
In truth, my best strength is in the writing anyways. I end up laying more weight on the words, and it reads more like a short story. It'll be more like a heavily illustrated pulp story. It helps the pulps are relatives in terms of media development.
I think the illustrator centrism of comics hurts comics a lot. So many of them are so badly written. And truely, this illustration centrism in comics makes them too resembling of movies. Words are images, no doubt, but a comic without words, or with minimal words, resembles storyboarding too much for my taste. More attention should be paid to when McCloud points out the broader history of comics (the comic on the Triumphal Arch, for instance, or the comics/language of the Egyptians or the ideographic tradition in the East, etc). As it stands, being a low-grade-static-movie only reinforces the popular view of the form being a bastard medium.
For instance, there should be more narrative voice in the stories, not just dialouge- and maybe even get rid of the balloons convention. If the dialouge is written right, you can tell who's saying what without needing the excessive attribution. I will admit this is very self-serving, for me, because my strength is in words, and not illustration. But the illustrators have been in control of the way the medium has developed too long.
On Control of Design
I think it's the control of how it looks that makes Flash so appealing. My biggest fear with Flash is that it may never become standard. i would like code I write now to be usable in a few years, and Flash isn't standard yet, and you end up having to learn scripting to make it really useful anyways- so javascript first:) Macromedia is really campaigning to get their work standardized into the browser, but I don't know they have the clout, nor do I think they will be the ones ultimately making that decision. The bigger fish may do away with them.
Bandwidth issues are really at the center of my wanting to use text rather than images, like you mention. Economy of the code and image files is exactly what I admire in you guy's comic. There are far too many comics out there that don't pay attention to that. Kinda like they want to ignore the producer voice in their head and go out all idealist. I can't read their comics, so I have no idea if their vision was worthwhile.
by William Beckerson - 12/04/2003 - 10:19
Here, let me pass this idea on to you:
The technology will catch up with E-Sheep eventually.