Flash 8

[dburdon]

I've been looking two Flash sites. Both are in version 8. The first has a whole lot of readable text in the Google cache. The other presents a blank screen.

I presume the software dvelopers kit in version 8 facilitates Google spidering? Any advice or pointers would be welcome.

[rms7design]

has introduced the ability to include metadata..... better spidering possibilities...

[dburdon]

RMS,

is there capacity in the meta data section to include a full page of text?

[Clarrie]

Originally Posted by dburdon

RMS,

is there capacity in the meta data section to include a full page of text?

Have you got the URLs of the two sites?

Is there more than one page of text? I'd be interested to see the code - I understand there's a rather dubious technique of adding spiderable text by placing the Flash in a layer overlaying the text. Spiders read the text and ignore the Flash, human visitors see the Flash, not the text.

Never actually seen a case of it, but I wonder if it might be something like that.

[weslinda]

Yeah, I've seen some discussions on that and the engines are not fond in the long run of that.

Would love to find out if there was a way that Flash is actually going to be text searchable in the future.

It would add a nice piece to allow more stylish and creative sites, but the limitations to it would be critical.

[dburdon]

Clarrie,

here's the URL of the site that appears to have been indexed by Google:

Flash: Google

Do you see anything dubious?

The other site's a potential client site. So I'll keep that under wraps.

[kgun]

Originally Posted by weslinda

Would love to find out if there was a way that Flash is actually going to be text searchable in the future.

I bought this book. My son experimented a little with making some flash animations.

What I learned from that was that a Flash animation is about:

1. Layers (to make animation).
2. Timelines (to make movement).

Then if you have layer on layer, which layer should the spider choose? What about the format?

Originally Posted by Clarrie

I understand there's a rather dubious technique of adding spiderable text by placing the Flash in a layer overlaying the text. Spiders read the text and ignore the Flash, human visitors see the Flash, not the text.

Any better solution than replacing the Flash with meaningful text as spiderfood?

[stevegreen]

Google isn't spidering the Flash content of this site, and it looks like they are doing a bit of black hat SEO.

You can see what's going on if you uninstall Flash and view the page. There is some text but what Google is picking up is the 'alt' text for a string of 1-pixel transparent images. These can be seen in the source code and they can be seen on screen by holding down the Tab key and watching the path of the focus (the images are links). You will see two series of white dots on the black background, one containing about 50 of these images and the other containing 20.

Each image has a huge amount of 'alt' text and it is clearly not intended for people to read because it is only displayed for a few seconds even if you were able to locate and hover over the 1-pixel image.

These techniques are easy for Google to detect and are exactly the sort of things I would expect them to penalise. I cannot think of any legitimate reason for designing the site this way.

Steve Green
Director
Test Partners Ltd / First Accessibility
http://www.testpartners.co.uk
http://www.accessibility.co.uk

[davebarnes]

I think the text the spiders do see is acutally a mistake caused by some bad code.

I went to http://www.webconfs.com/search-engin...-simulator.php and entered http://www.originalhouse.es/ as the URL to "spider".

I then looked at the HTML source code for the page and it appears that a coding error causes some text to be seen by the spider simulator. I think that correct coding would actually cause no text to be seen.

,dave

[kgun]

Think of the following.

Exactly the same text on the flash image as in the alt tag.

[Faglork]

Originally Posted by stevegreen

Google isn't spidering the Flash content of this site, and it looks like they are doing a bit of black hat SEO.

A "bit"? Just look at the source code - they have a whopping 64k (!!!) in ALT texts. IMO just another example that proves the inefficiency of Google's counter-spam measures, if there are any at all.

Surprisingly, the only legit place to put text in - the <noscript> tag - is comparably empty.

Maybe they just don't know it any better.


Cheers,
faglork

[dburdon]

Faglork,

I know you're against alt text - so am I apart from legitmate use. But do you know how the Software Developer's Kit works in Flash 8?

[bhorsting]

You are probably referring to the Flash Search Engine SDK (as opposed to the Flash Player SDK, which is the Flash Player source code, granted to you by Adobe if you wish to port Flash Player to a new platform), which basically is a series of command-line tools to extract texts from SWF files (which will only work if these strings actually EXIST in a flash file, which in the case of anything more advanced than a simple animation with text will not be the case).

That this website is being crawled is not a mistake nor a coding error, but an intentional attempt to escape the invisibility of Flash textual content.
1-pixel linked images with ALT tags are used to not visually interfere with the rest of the page.

The problem is that the ALT descriptions are waaay too long at the moment; the client can manage these and will be asked to tone them down.

[dburdon]

Bhorsting,

do you have inside knowledge of the project and processes involved?

Is this a legitimate, i.e. Google approved, use of the SDK for SEO purposes?

[kgun]

It may be possible to scan a picture and put the scan in a bitmap.

Keywords:

ocr scan

ocr scan to bitmap

OmniPage Professional

Link:

An introduction to scanning.

Optical character recognition

Artificial neural network

Pattern recognition

Then it should be possible to "read" the text on the picture.

(When) is it possible for a SEbot to do it?