Creative Commons license icon

Reply to comment

Thanks for taking the time! Here's the revisions I made (original, revised), and explanations in rough order of importance:

Images should be offset from the text with margins

We don't use much white space at Flayrah, but a little is a good thing. For right-aligned images, style=margin-left:5px is good. For left-align, I've been moving towards margin-right:8px. Preview and see how it looks. If there's an awkward word hanging at the end of a line, replace the space after it with   (non-breaking space) to move it to the next.

You need to set Input format to Full HTML for margins (or any styling) to work. It's not there by default for security purposes.

This is also the point to select a cutoff for the lede ("Split summary at cursor"), and to adjust the size of images, while preserving good flow. Consider shifting them up a line, and/or using top/bottom margins to adjust for awkward gaps.

Images should be thumbnailed, not just resized

With images, we have a few conflicting issues:

  1. We want the images to be small enough to fit on screen and not crowd text (but large enough to see)
  2. We want the page to load fast/not use much data (our bandwidth is a minor issue, but some are on slow mobile connections) - also, Google priorities faster pages
  3. We want users to be able to see the highest-quality version of the image at full size if they want
  4. We want other sites and search engines to read our page, recognize we have larger images, and use them

Resizing an image by setting the width and/or height only solves the first of these. Instead, we create a thumbnail of the image at a lower resolution - solving 1 and 2 - and link that to the full-size image to meet 3 and 4. You can still use an image tag with width to figure out what size to use, but be sure you don't forget to replace the src after creating the thumbnail.

You can create thumbnails with the resize button in the upload dialog - leave "create a new image" checked, enter the width you decided, leave height blank, and tab to height to have it auto-fill. Or, resize the original on your computer and upload it as well. The latter becomes necessary if you also want to crop the image to remove cruft or highlight a particular section.

Both width and height should be set explicitly on images

Not setting both attributes mean browsers have to load the image and then reflow the page afterwards. It also makes it hard for bots (and Flayrah itself) to figure out what images to use in a summary.

If you're using thumbnails as above, this shouldn't be a problem; the image dialog shows both. If you have to use the original (say, it's a small change and the source is highly compressed), you can still use the dialog, just don't create the thumbnail.

There should ideally be at least one image with a display size greater than 280x150

The largest image image will be used on Twitter as part of a large image card. Otherwise, you get a small square image next to the title, which usually doesn't look as good. The algorithm picks the largest image based on width X height as set in the image tag. If the image is linked, it'll use the linked image (but doesn't know its size; a big picture with a small thumb won't be picked).

The largest image is not always the first image in the story - it can be hard to fit a large image into the lede. In dronon's recent review, it picked the last image. In this case, I felt the abstract cover was a central image and beneficial to display at a high resolution due to its detail, so I chose to increase its size and so caused the algorithm to pick it.

If no images are this big, but there are YouTube or Vimeo links/embeds, it'll use the first of those.

Images should have descriptive alt attribute text (and title attribute if necessary)

The alt tag provides information about an image for those who cannot see it or who have chosen not to load it. This includes bot like Google Image Search. We want people looking for images to find them on Flayrah because they may be interested in the story.

If there's no need for a visible caption, but you have a comment for the curious to explain the image, you can use the title attribute. This'll popup on mouseover in most browsers. In this case they are all self-explanatory, so no titles.

(Having a visible caption means using a div with float:left/right and width attributes, rather than aligning the image. You can then add a newline inside the div and add italicised caption text. I've added an example to the guidelines.)

Amazon.com links should have ?tag=flayrah-20 at the end; all products should be linked

This means we get money! It's less than $10 a month, but it helps with the server bills. You can shorten it to something like http://www.amazon.com/dp/0441240860?tag=flayrah-20 (just check that it loads).

An author may not provide links for all works. Here I added one for Treaty Planet. (I also linked the cover references.)

Works should be announced on Twitter when published

Tweets don't just happen for regular stories like they do for newsbytes - you have to tick the Tweet box when publishing.

Page numbers on quotes (including the parentheses around them) should be enclosed in <small>

This is just a stylistic thing. You may also need to use &nbsp; in place of a regular space in-between "p./pgs." and the page number if it's near the end of a line, otherwise you can end up with the number on a new line by itself.

If this doesn't show up, you probably didn't select "Full HTML" input type yet.

There should be a single space after image tags (or their links) when followed by text

This is because the alt text might be turned into actual text (especially by bots), in which case it could run up against the following text. For example, you might get "Crisis at DoonaThis book is..." in a Google News description.

Images references and links to Flayrah should not use scheme ("http:" or "https:"), but start with //

Flayrah can be accessed via https: as well as http:. https: is slightly slower, but use of http: means people monitoring a network (say, at work) can see both the addresses of the pages and images you are accessing, the content itself, and your login credentials.

If a secure connection is maintained, they can only see the server IP address, which may be shared by other sites, like WikiFur. (Although, the page will also be in your browser history; it's best to assume work computers are fully monitored.)

If you have a http: image on a https: page, either the browser won't load the image, or it leaks the file accessed to watchers. Either way, it'll complain. Take out the scheme, and it'll use whatever the user accessed the page with. Same for links.

This also applies to links (and embeds) from WikiFur, Wikipedia, YouTube, Vimeo, and other sites accessible both ways.

HTML character entities such as &lsquo;/&rdquo; and &ndash; should be replaced with the actual characters

This doesn't affect the visible output, but it's slightly faster for readers to load and makes it easier to find things that need editing. Personally, I preview the page and then copy the story text to notepad, using successive find-and-replaces to convert each entity (I'm already doing this to convert http://www.flayrah.com to //www.flayrah.com), but a tool might be faster.

Now - is it reasonable to expect one person to do all this? Perhaps not. Sometimes I only get partway. You can always do what you have time for, and leave the rest to another editor. Whether you publish the article in the interim depends on its urgency. While the issues above are mostly stylistic, a story should not be published without copyediting or fact-checking - and features often require substantive editing as well.

Reply

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <img> <b> <i> <s> <blockquote> <ul> <ol> <li> <table> <tr> <td> <th> <sub> <sup> <object> <embed> <h1> <h2> <h3> <h4> <h5> <h6> <dl> <dt> <dd> <param> <center> <strong> <q> <cite> <code> <em>
  • Lines and paragraphs break automatically.

More information about formatting options

CAPTCHA
This test is to prevent automated spam submissions.
Leave empty.