Diperbarui 30 Agustus 2010 oleh Dani Iswara
Better than usual podcasts, they claimed it as podiocasts. How accessible is Odiogo result? This review is based on my opinion as a user. I have no Odiogo’s account yet. So, with the help of Cahya Legawa (cahya.legawa.com), using one of my blog post, we did a test page. My post titled Semantic XHTML Format Reference reposted with my permission on his blog post using the same post title. Odiogo’s text-to-speech (TTS) will read the content/post only. As similar as delivered by feeds syndication. The blog post/feed needs at least 24 hours to be converted by Odiogo server. You may download its podcast result (at podcasts.odiogo.com; as MP3 format; 3MB in size).
Result and suggestions
My comment for the result:
- The speech is good & clear. At least, better than Adobe Reader’s text-to-speech machine.
- It doesn’t read the title attribute.
- The left and right parenthesis will not read at all.
- Words in capital letters will be read as an abbreviation, read aloud per letter.
- Headings and subheadings are not distinct, no pause after it.
alttext is missing, and
altwill be read as "alt".
- Links are not identified.
- No pause on
- No emphasize while reading
- This form for example,
<code>, will be read as "and lt code".
- Then the closed element above,
</code>, will be read as "and lt slash code".
Of course, Odiogo and Adobe Reader text-to-speech softwares are not using screen reader algorithm. For better accessibility experience, use a real screen reader. Or try those text-to-speech tools to read the Fangs—screen reader emulator—results. See how Fangs read Twitter homepage (Dani Iswara .com).
My suggestions for Web/blog authors to optimize Odiogo result:
- Expands all abbreviation if needed.
- Use capital letters for abbreviations only.
- Puts break punctuations in every listings.
- Headings need break punctuations also. But there is no rules to put it.
- Use active sentences.
- Use short and clear words.
Read also how to make Web content more readable (Dani Iswara .com).
My recommendation for Odiogo algorithm, compared to Fangs:
- Links should be distinct. Every links should be read as "links [followed by anchor text]".
- Semantic elements should be read properly. Users should know that the text is a headings, code, quotation, caption, table row/column, etc.
- Emphasize elements have a higher pitch.
- There should be a pause in such as headings, subheadings, listings, and definition lists.
They may include more features for some applications (blog.odiogo.com) also. Can we see a better accessible tool of Odiogo?