Declaration count issues

[j9t]

I’ve taken some declaration count samples with CSS Stats and they looked good, however I’ve also found some rather severe outliers.

Technorati, for example, is reported to use 111 declarations, 75 of them unique: http://cssstats.com/stats?ua=Browser%20Default&url=http://technorati.com/#unique-totals.

However, Technorati uses at least 2 style sheets of which each has more than those 111 declarations alone (like http://technorati.com/wp-content/plugins/revslider/rs-plugin/css/settings.css?ver=4.0 and http://technorati.com/wp-content/plugins/ultimate-social-deux/public/assets/css/style.css?ver=3.1.6).

The numbers are hence completely off.

On scanning I couldn’t make out the issue—could you investigate (and fix)?

(As I’m working on a little study that uses CSS Stats for such counts, how great would you believe CSS Stats’ error rate to be for declaration counts?)

[mrmrs]

Hi @j9t - it seems your css isn't valid and has a number of parse errors. Regrettably, we have not had the resources to provide better error handling in the UI for things like this. Our apologies for causing confusion at all.

If you provide valid css, there shouldn't be any cases (I know of) that will provide incorrect declaration counts.

Guessing it is either choking on the ms-filter syntax you are using or this syntax error in a comment close to the bottom of the file (or maybe both)

I don't believe the error rate is high, or existent on valid css. Don't have info on how many requests fail due to bad css but we might be able to look into this at a later date.

[mrmrs]

Hi @j9t - it seems your css isn't valid and has a number of parse errors. Regrettably, we have not had the resources to provide better error handling in the UI for things like this. Our apologies for causing confusion at all.

If you provide valid css, there shouldn't be any cases (I know of) that will provide incorrect declaration counts.

Guessing it is either choking on the ms-filter syntax you are using or this syntax error in a comment close to the bottom of the file (or maybe both)

I don't believe the error rate is high, or existent on valid css. Don't have info on how many requests fail due to bad css but we might be able to look into this at a later date.

[j9t]

Ah, interesting. Yes, as you normally throw errors when invalid CSS is fed into CSS Stats, that would help in instances like this, too.

Note that I’m not associated with Technorati; this was an example.

[j9t]

Ah, interesting. Yes, as you normally throw errors when invalid CSS is fed into CSS Stats, that would help in instances like this, too.

Note that I’m not associated with Technorati; this was an example.