[BillP]

I'm testing this compressor, and I'm having a very hard time reconciling the statistics generated by the compressor (as shown to me on the bottom of every page) to the real size of the page.

I am using FireFox with the Web Developer Toolbar, which offers a quick summary of the space used on a page.

The document itself (the HTML for that page, less all external data such as images, CSS, etc.) is the only thing that gets compressed, and that's correct, but I am seeing around 10% savings, but the stats consistently say 22% or so savings.

I'll keep staring at the calculations, but I think the stats are not accurate.

I wonder if Firefox is measuring the GZipped output? My server uses gzip everywhere, and the stats are definitely calculated pre-gzip, the difference could be in the exploded/deflated sizes.

I don't see that I'll be able to measure the true bandwidth savings, as 10% on a page is meaningless compared to the images and CSS that go along with it. For example, my forums home page has 9K of data, 46K of images, 1K of style sheets, and 13K of scripts (vbulletin_global.js, for example). The stats say that my page compression was: [page compression: 65.53 k/84.13 k (22.11%)]

If I turn off the page compression, the document increases from 9K to 10K, all other stuff remains the same, so Firefox reports a 71K page.

I just can't reconcile this. Can anyone else see similar results?