# Finding out what's in the \<head\>?

**URL:** <https://discuss.httparchive.org/t/finding-out-whats-in-the-head/2119>\
**Category:** Analysis\
**Created:** [March 7, 2021, 12:05pm UTC](https://discuss.httparchive.org/t/finding-out-whats-in-the-head/2119 "2021-03-07T12:05:09Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Nathaniel](https://yyz1.discourse-cdn.com/flex035/user_avatar/discuss.httparchive.org/nathaniel/32/1848_2.png) [@Nathaniel](https://discuss.httparchive.org/u/Nathaniel)\
**Post date:** [March 7, 2021, 12:05pm UTC](https://discuss.httparchive.org/t/finding-out-whats-in-the-head/2119/1 "2021-03-07T12:05:10Z")

</div>

Hello I’m new. Recently I started making a website with a strict page weight of `30kB`. One page was over budget, so I started looking for things I could safely remove from the page.

I was delighted to find out [from the almanac](https://almanac.httparchive.org/en/2020/markup#favicons) that `<link rel="icon" href="/favicon.ico">` is redundant as browsers look at the `domain root` for `favicon.ico`.

This got me looking at all the other things I and others stick in their `<head>`, and I’ve since found that most of it is either totally useless, or has a work around that allows you to remove the tag.

I’m writing up all my findings in an article hoping that people will remove these tags from their sites — I know these optimizations are tiny in comparison to what can be achieved with images and video — but it’s something.

I’d love to have some statistics on what `<meta>` and `<link>` tags people are using. For instance, how many sites are including:

```sql
<meta name="HandheldFriendly" content="True">

```

A tag that tells the `AvantGo` browser that you’ve optimized your website to work on a `Palm Pilot`!

Anyway, I’d love some advice on how to get this infomation. I’m going to start following the [Getting Started Guide](https://github.com/HTTPArchive/httparchive.org/blob/master/docs/gettingstarted_bigquery.md) and finally get round to learning `SQL`.

Thanks, i’ve really enjoyed reading back on older posts, really interesting stuff.

---

<div class="post-metadata">

**Author:** ![rviscomi](https://yyz1.discourse-cdn.com/flex035/user_avatar/discuss.httparchive.org/rviscomi/32/835_2.png) [@rviscomi](https://discuss.httparchive.org/u/rviscomi)\
**Post date:** [April 20, 2021, 8:27pm UTC](https://discuss.httparchive.org/t/finding-out-whats-in-the-head/2119/2 "2021-04-20T20:27:13Z")

</div>

Hi Nathaniel,

I think the best option would be to leverage [custom metrics](https://github.com/HTTPArchive/legacy.httparchive.org/blob/master/custom_metrics/README.md) that parse the DOM at runtime and emit stats/data that we can analyze later in BigQuery.

For example, we already have a custom metric that reports on all meta tags in the document head:

> <https://github.com/HTTPArchive/legacy.httparchive.org/blob/6ac1244d7fe39bea6b10f98f3890b4d4360d11a0/custom_metrics/almanac.js#L221-L226>

You could query that metric to get info about your `HandheldFriendly` example.

I’d suggest making a list of all of the things you’d want to analyze in the head and we can see which are already available as custom metrics, and implement the rest.

Rick