5 ms·

I managed to get just over half-way and recall 58 of them before I ran out. I find it interesting which are omitted. I assume these tags are just fully removed

by jchw 2y ago

I managed to get just over half-way and recall 58 of them before I ran out. I find it interesting which are omitted. I assume these tags are just fully removed from HTML5, although most of them continue to work just fine with the exception of <blink>. - <blink> - <marquee> - <tt> - <center> - <frame> - <frameset> That said, it accepts both <object> and <embed>, which I found moderately surprising, since I thought that those were mostly only used for ActiveX and NPAPI plugins. Maybe the last major round of deprecations for HTML came just a tad too early to kill them off, or maybe there is more to them than that. Some tags I feel are underrated in handwriting HTML that are still in the spec: <figure> and <figcaption>, good for semantically representing images with captions. <thead>/<tbody>/<tfoot>, semantically segments a table; In theory could be used by assistive technology, or possibly renderers that need to paginate HTML and want the table headers to repeat on each page (does any agent do this?). <picture> + <source>, pure-HTML method of having responsive images in an HTML page, can support a fallback for browsers that don't support it, too. On that note, the srcset and loading attributes of img are very useful too, srcset as an alternative to <picture>, ostencibly aimed at use cases where the different sources are identical aside from pixel density (vs <picture> which is meant to contain potentially more significantly different images for different responsive layouts.) I wonder which things I missed. That would be a nice feature to add. (I know I missed a few table elements because I just thought of colgroup while typing this.)

mbo 2y ago
No love for <q>? It does curly quoting!
- jchw 2y ago
  Wow, that's awesome! I've been remembering the HTML entities for double quotes for so long, I actually had no idea this tag existed. What's even more fascinating about <q> is that it is localized too! Wow! Check it out. If you do... <p lang="en"><q>My English quote</q></p> you will of course get... “My English quote” But, if you do... <p lang="ja"><q>日本語の引用</q></p> Then you get: 「日本語の引用」 Magic! There's only one real problem I have with this: it seems to be implemented on the layout-level using ::before and ::after pseudo-elements, and thus the quotation marks are intangible. That's a real bummer, which has the result of making the quotation marks not able to be copied or selected (and their visibility to screen readers may also be impacted?) So unfortunately, for now, I am going to continue using good ol' ldquo and rdquo. But, that's pretty neat.
bsd_source 2y ago
It also doesn't accept keygen (Although that never really worked.)
- anderskaseorg 2y ago
  It worked in all major browsers until it was removed around 2017–2019. I personally used it, as part of MIT’s authentication systems.
jefftk 2y ago
It also rejected one of my favorites, the long-deprecated <xmp>
- myfonj 2y ago
  <xmp> and <plaintext> never die! I think/hope/believe that even though these elements fell out of grace of the standard bodies [1] long time ago, implementers simply cannot go against backwards compatibility here and remove them like spec suggest and MDN scaremongers that it "was not implemented in a consistent way" [2], yet last time I checked, the support in major browsers was pretty solid [3]. - [1] yet https://html.spec.whatwg.org/multipage/obsolete.html#xmp https://html.spec.whatwg.org/multipage/obsolete.html#xmp tells still acknowledges that "User agents must treat xmp elements in a manner equivalent to pre elements in terms of semantics and for purposes of rendering. (The parser has special behavior for this element though.)" - [2] https://developer.mozilla.org/en-US/docs/Web/HTML/Element/xmp https://developer.mozilla.org/en-US/docs/Web/HTML/Element/xm... - [3] https://github.com/mdn/browser-compat-data/issues/2168#issuecomment-393708922 https://github.com/mdn/browser-compat-data/issues/2168#issue...
  - jefftk 2y ago
    I do actually use <xmp> when authoring blog posts, though I wrote a preprocessor to convert to escaped <pre> in case some rendering environments don't support it: https://www.jefftk.com/p/i-want-xmp-but-i-know-why-i-cant-have-it https://www.jefftk.com/p/i-want-xmp-but-i-know-why-i-cant-ha...
assimpleaspossi 2y ago
Both blink and marquee have never been part of any HTML specification since the dawn of HTML. But marquee was inserted into the current version for the sole purpose of marking it obsolete because it was in such widespread use.
- a_t48 2y ago
  Fun fact - the website used for submitting visas to come to India still uses marquee.
sdsd 2y ago
I actually used marquee on two recent sites I made (for fun). On darigo.su if you open Napster (in the pokeball button computer), and on hnchan.darigo.su. I love marquee.
- jchw 2y ago
  > hnchan.darigo.su > NB: we've left a SQLi vuln in this site. If you're smart enough to find it, congrats: you're a mod. That's certainly a creative approach.
hobofan 2y ago
I think for some fallback mechanisms <object> is the only way to implement them in pure HTML. E.g. I once built a logo element, where the logo may be either at logo_A.jpg, logo_B.jpg or logo_C.jpg. As there wasn't any API to determine which of them would exist in advance, I instead used three nested <object> tags. They would request the image in sequence, and the first one that would respond with an image is displayed.

by jchw 2y ago

mbo 2y ago

No love for <q>? It does curly quoting!

jchw 2y ago
Wow, that's awesome! I've been remembering the HTML entities for double quotes for so long, I actually had no idea this tag existed. What's even more fascinating about <q> is that it is localized too! Wow! Check it out. If you do... <p lang="en"><q>My English quote</q></p> you will of course get... “My English quote” But, if you do... <p lang="ja"><q>日本語の引用</q></p> Then you get: 「日本語の引用」 Magic! There's only one real problem I have with this: it seems to be implemented on the layout-level using ::before and ::after pseudo-elements, and thus the quotation marks are intangible. That's a real bummer, which has the result of making the quotation marks not able to be copied or selected (and their visibility to screen readers may also be impacted?) So unfortunately, for now, I am going to continue using good ol' ldquo and rdquo. But, that's pretty neat.

bsd_source 2y ago

It also doesn't accept keygen (Although that never really worked.)

anderskaseorg 2y ago
It worked in all major browsers until it was removed around 2017–2019. I personally used it, as part of MIT’s authentication systems.

jefftk 2y ago

It also rejected one of my favorites, the long-deprecated <xmp>

myfonj 2y ago
<xmp> and <plaintext> never die! I think/hope/believe that even though these elements fell out of grace of the standard bodies [1] long time ago, implementers simply cannot go against backwards compatibility here and remove them like spec suggest and MDN scaremongers that it "was not implemented in a consistent way" [2], yet last time I checked, the support in major browsers was pretty solid [3]. - [1] yet https://html.spec.whatwg.org/multipage/obsolete.html#xmp https://html.spec.whatwg.org/multipage/obsolete.html#xmp tells still acknowledges that "User agents must treat xmp elements in a manner equivalent to pre elements in terms of semantics and for purposes of rendering. (The parser has special behavior for this element though.)" - [2] https://developer.mozilla.org/en-US/docs/Web/HTML/Element/xmp https://developer.mozilla.org/en-US/docs/Web/HTML/Element/xm... - [3] https://github.com/mdn/browser-compat-data/issues/2168#issuecomment-393708922 https://github.com/mdn/browser-compat-data/issues/2168#issue...
- jefftk 2y ago
  I do actually use <xmp> when authoring blog posts, though I wrote a preprocessor to convert to escaped <pre> in case some rendering environments don't support it: https://www.jefftk.com/p/i-want-xmp-but-i-know-why-i-cant-have-it https://www.jefftk.com/p/i-want-xmp-but-i-know-why-i-cant-ha...

assimpleaspossi 2y ago

Both blink and marquee have never been part of any HTML specification since the dawn of HTML. But marquee was inserted into the current version for the sole purpose of marking it obsolete because it was in such widespread use.

a_t48 2y ago
Fun fact - the website used for submitting visas to come to India still uses marquee.

sdsd 2y ago

I actually used marquee on two recent sites I made (for fun). On darigo.su if you open Napster (in the pokeball button computer), and on hnchan.darigo.su. I love marquee.

jchw 2y ago
> hnchan.darigo.su > NB: we've left a SQLi vuln in this site. If you're smart enough to find it, congrats: you're a mod. That's certainly a creative approach.

hobofan 2y ago

I think for some fallback mechanisms <object> is the only way to implement them in pure HTML. E.g. I once built a logo element, where the logo may be either at logo_A.jpg, logo_B.jpg or logo_C.jpg. As there wasn't any API to determine which of them would exist in advance, I instead used three nested <object> tags. They would request the image in sequence, and the first one that would respond with an image is displayed.