Last modified : 2014. 10. 13


This page describe the abstract policy of KARAS which is different from design policy.

Concept and Rationale Should be Shown

Concept and rationale of open source product must be shown, even if it is too small product. There are several reasons why.

One is to prevent the similar products proliferation. This is same as "License Proliferation". In order to prevent the proliferation, it is necessary to indicate the concept and rationale, and get the many users approval. Further, by indicating the rationale, it becomes possible to discuss more specific about the design.

Another reason is in order to generate a more better product in future. The good thing about the product can be employed in a new product. Because computer get to be able to handle more things in future, it is clear that the current product get to be unable to handle all of it. In order to make a new product, it is important to know the history of the development of the same type products. And, repeating same mistake is not good.

A Rationale to be Indicated

As WHATWG described in Is design rationale documented? , indicating all of the discussion, examination, and rationale is too hard, and such document will consist of a lot of pages. This is certainly true. However, at least, important reason to determine the design should be shown.

Markdown and reStructuredText are Excellent, but not Enough

Markdown and reStructuredText are excellent invention. Especially, because Markdown is using popular markup in a day-to-day habit, it is very nice as an interface. However, to me, Markdown is not enough, and reStructuredText is too complicated. I also checked another famous lightweight markup languages such as textlie, but those were not able to convince me. I found that only SkrivML has been designed in similar concept and philosophy, still, I was not satisfied. I believed that my idea is splendid, and developed KARAS.

In many case, Markdown is enough. I think so. And I think the greatest point of Markdown is the name.


This is a list of LML and PDL (Page Description Language). Several pages is a Japanese page. And related information has been well organized in the Wikipedia - Lightweight markup language.

Reference other than LML

In order to design the syntax of KARAS, to refer to LMLs is obvious. KARAS also refers to programming languages and regular expression.

And, I considered what kind of text and document could be written by using KARAS. I checked academic papers, news papers, packages of medicine and snack, direct mail, novel, dictionary and others. Except for the difficult pattern that to be lay out by using HTML and CSS, KARAS is able to write many kinds of documents.

Possibly, functions of the KARAS is not enough. But if there are more functions, it becomes more complex. When need such wide range system, we should use HTML without creating a new language. And if needed, KARAS is able to extend the system by using Plugin syntax.

Design concept such as Semantic Web should be Eliminated in the Distant Future

Concept of the Semantic Web will remain for a long time. However, I think automatically identifying and classifying the information or property that shown in a text, is ultimate goal of this field. And I hope it is realized in the future.

Such system user will not have to enter the meaning of the text like LML does, and can write the document more naturally than now. I wish KARAS is possible to use until such future will come.

However, there is no change that human detect and utilize the document with their eyes. Therefore, markup technique to make the text and the block clearly will not disappear for a while. Also format to provide meaningful text such as HTML will never gone, in order to save the result of analysis of document.

HTML5 impairing the Value

Now we can see a lot of articles of HTML5 (in 2014). However, in truth, most of the contents show about canvas element or combination of Javascript and CSS3. Newly defined element has been neglected. But I know the feeling well. Hit the tag with thinking the meaning of the text is very bothersome. KARAS will beat the bothersome and increase the value of HTML.

Need not be limited to HTML

KARAS has an HTML as standard output format, but not only this, you can employed KARAS syntax in any environment. When you need EPUB or PDF, regard the HTML as intermediate format, to convert by existing tool is one of idea. Of course, developing the dedicated converter is also good. You can write the parameter that important to EPUB or PDF with plugin syntax of KARAS if you need.

My thinking about HTML

I would like to mention my thinking about HTML in here. These were considered when designing the KARAS. My basic stance is this. To be simple, to be easy, and mistakes should be fixed. This is different from W3C or any others.

Please note the following things. I have only read a part of the document which shows record of the discussion WHATWG and W3C has done. I don't know about all of the history or contributor's thinking. And, I have a gratitude and respect to the contributors.

small Element should be Renamed

In 2014 Jan., In HTML5, small element is defined that it should be used to mention disclaimer or note. I think this design is wrong. Because we can not guess this rule from the name of small.

In order to write small text, we often use a word small. In generally, to make a part of text small, we will use <span class= "small">. When I thought about this, I felt that the design of small element is clearly wrong. Even if it is still used in a history.

cite Element should be used to Indicate a Source

In 2014 Jan., In HTML5, cite element is defined that it should be used to indicate title of work such as book, paper, or any other. I think this design is wrong. Because original meaning of word cite is to quote. When we quote something, either is fine if it is speech of a person or something work. However, based on the long lasted (bad) habit, we can not use cite element to indicate a speaker, even it can use to indicate a title. This is very disappointing. Define the rules on the basis of the habits or de facto standard is not bad. However, We must not forget that we can correct mistakes.

I hope that KARAS user use cite element to indicate a speaker. In KARAS, text enclosed with @@ is outputted as cite element. In addition, to increase one more @, it is outputted as small element.

If you approve of my thinking, please use cite element to indicate a speaker, with writing two @ mark. If you think that we should take a W3C or WHATWG opinion, you can output small element with writing three @ mark. Or, you should use strong(b) element or span element. There is a syntax called inline-group to output span element in KARAS.

In 2014 Jan., speaker of Twitter-timeline is markedup with strong element. On the other hand, in a Blockquote article of Twitter-Bootstrap, speaker is markedup with small element.