Question about the search function

[PhiloVance]

Does the global search function include searching through attachments? Does it include the title of the attachment?

Forgot to mention this is the web based version.

[n5bb]

A basic search looks through the raw message content, as far as I can tell. This means it will find the file names of attachments and the content of text attachments (such as .eml emails or .txt text files). The content of binary files (such as PDF documents) can't be searched.

Bill

[odedp]

Originally Posted by n5bb

A basic search looks through the raw message content, as far as I can tell. This means it will find the file names of attachments and the content of text attachments (such as .eml emails or .txt text files). The content of binary files (such as PDF documents) can't be searched.

Bill

Why?
Evernoe, for example, searches PDF documents (and more) without any problem

[BritTim]

Originally Posted by odedp

Why?
Evernoe, for example, searches PDF documents (and more) without any problem

It would be technically possible, but extremely difficult, to search binary attachments for text.Let us take an example of what might be involved:

• Assume one attachment is a 7-zip archive containing a Word document and a PDF file
• First, the search funtion will need to correctly determine the type of the attachment (non trivial in itself: relying on the attachment name is error prone and even file signatures are not fool proof)
• Next, each individual member of the archive needs to be individually extracted
• The type of the Word file must be correctly deduced, and a type specific function used to search for text inside it.
• Ditto for the PDF file

I have no expectation that FM will enhance the Cyrus search facility to add this capability any time soon.

[odedp]

Originally Posted by BritTim

It would be technically possible, but extremely difficult, to search binary attachments for text.Let us take an example of what might be involved:

• Assume one attachment is a 7-zip archive containing a Word document and a PDF file
• First, the search funtion will need to correctly determine the type of the attachment (non trivial in itself: relying on the attachment name is error prone and even file signatures are not fool proof)
• Next, each individual member of the archive needs to be individually extracted
• The type of the Word file must be correctly deduced, and a type specific function used to search for text inside it.
• Ditto for the PDF file

I have no expectation that FM will enhance the Cyrus search facility to add this capability any time soon.

Thank you BritTim

[n5bb]

Originally Posted by odedp

...Evernoe, for example, searches PDF documents (and more) without any problem...

Did you mean "Evernote"? That's not an email service, and you have to be a premium Evernote paid user to search PDF's which don't have embedded text content. Some PDF files have been scanned by OCR (optical character recognition) software and the text results (which may have errors) are appended to the PDF file, allowing it to be searched for text content. Evernote does this automatically for premium accounts (only), and it will cost you US $5/month or $45/year.

As BritTim points out, searching the content of attachments is costly, especially if they are not already in a plain text format. But it's possible to provide this service in many cases, and as computing and storage costs continue to drop I would imaging that such features might be available for at least the more expensive email accounts in the future. Just as with an anti-virus scanner, it should be possible to scan each document in the background and add plain text so it could be later scanned. But this text would have to be stored in some way to associate it with each message, and all of this needs to be highly secure. I don't think this is high on Fastmail's priority list.

Bill

[odedp]

Originally Posted by n5bb

Did you mean "Evernote"? That's not an email service, and you have to be a premium Evernote paid user to search PDF's which don't have embedded text content. Some PDF files have been scanned by OCR (optical character recognition) software and the text results (which may have errors) are appended to the PDF file, allowing it to be searched for text content. Evernote does this automatically for premium accounts (only), and it will cost you US $5/month or $45/year.

As BritTim points out, searching the content of attachments is costly, especially if they are not already in a plain text format. But it's possible to provide this service in many cases, and as computing and storage costs continue to drop I would imaging that such features might be available for at least the more expensive email accounts in the future. Just as with an anti-virus scanner, it should be possible to scan each document in the background and add plain text so it could be later scanned. But this text would have to be stored in some way to associate it with each message, and all of this needs to be highly secure. I don't think this is high on Fastmail's priority list.

Bill

Yes Bill, I meant "Evernote"...
I am a premium Evernote user and search many of my stored PDF files frequently.
But I am also a Fastmail premium user (Enhanced account) and think that searching the content of PDF attachments would be a "nice to have" feature.
Thank you for your detailed (as always...) explanation