creating thumbnails without downloading entire image.

Is it possible to create a solution that creates dynamic thumbnails
from iamges without downloading the entire image.

This is on this newsgroup because its primarily a networking question.
Can I request each image seperately and in the order that I want. In
particular I know that sometimes when you download the image it goes
from top to bottom, other times it goes from blurry to sharp. If I
can request the order of information transfer, I could for each image
just get the blurry versions and use that information to generate
sharp small thumbs. These will not be ideal (ie average of
surrounding pixels) but this will be fine. Thank you very much for
any information.

If this is possible is there an implementation, or do you guys have
some tips as to how I could go about implementing this; I imagine it
would be through some sort of hack in the TCP request, but I'm not
well versed enough in these methods to know.

-Ashot Petrosian
UTCS

Ashot Petrosian wrote:
> Is it possible to create a solution that creates dynamic thumbnails
> from iamges without downloading the entire image.
>
> This is on this newsgroup because its primarily a networking question.

It's not. - Read on.


> Can I request each image seperately and in the order that I want. In
> particular I know that sometimes when you download the image it goes
> from top to bottom, other times it goes from blurry to sharp. If I
> can request the order of information transfer, I could for each image
> just get the blurry versions and use that information to generate
> sharp small thumbs. These will not be ideal (ie average of
> surrounding pixels) but this will be fine. Thank you very much for
> any information.

Your question is about HTML and its techniques, this NG's primary
focus is on technical questions related to networking. HTTP, which
transports HTML, is in the application level, which nobody here
cares much about.

But to answer Your questions anyways: You can request each image
seperately. You will first ask for the HTML source of any given
page You want to view. H_T_ML is text only, and may contain refe-
rences to images that Your browser shall display within that text
area. Your browser (if not lynx) will recognize those references
and request each of those images in turn from the server. - So,
You need to get the page source first, parse it for image references
and then, ask for the images to be sent. That is what every browser
does.

When an image goes "from blurry to sharp", there are two concepts
for that: If it's ".gif", there's an "interlace" functionality in
this format (gif87a and gif89) which controls this. Non-interlace
will transfer the image line by line, while interlace will send
like $LINE_0; $LINE_N; $LINE_(N*2) .. $LINE_(N*n), and then start
from the top with $LINE_(0+1); $LINE_(N+1); $LINE_(N*2+1) and so on.
There's a similar thing for ".jpeg", but I'm no expert on that.
In HTML, there's also a "lowres" parameter for the <IMG> tag, where
You can specify a source that contains a smaller version (bytewise,
that is; that has nothing to do with the size of the displayed img)
of the to-be-displayed picture (like lower resolution and/or black-
/white) that will be loaded and displayed first, and only after all
such placeholders for the pictures of one page are loaded, the actual
images are being transmitted.

Only in this latter case, where "lowres" is used in an HTML <IMG>
tag, You can save bandwith by only downloading the lowres source;
with the interlace mechanism, You will have to load the entire file
first. - You can later resize it to save disk space, but to be able
to do so, You need to have the entire image.


> If this is possible is there an implementation, or do you guys have
> some tips as to how I could go about implementing this; I imagine it
> would be through some sort of hack in the TCP request, but I'm not
> well versed enough in these methods to know.

Again, You want to talk HTTP. That's applications. Go find an appro-
priate group for that. - Here, it's "FTP-County", with some more low-
level speakers. We don't understand Your high-level conversation...

Just kidding, Ashot, but I hope I got You on the right track.


Cheers, Jack.

--
----------------------------------------------------------------------
My personal reading of the string "MicroSoft" expands to "NanoWeak"...

Ashot Petrosian <ashot@molsoft.com> wrote:
> Is it possible to create a solution that creates dynamic thumbnails
> from iamges without downloading the entire image.

No. If it's straight bitmap, then, I guess, you can request every other
1k byte of the image to reduce the bandwidth. But, most images are
compressed (ie. GIF/JPEG/PNG), so you won't be able to uncompress due to
"missing" data.

> If this is possible is there an implementation, or do you guys have
> some tips as to how I could go about implementing this; I imagine it
> would be through some sort of hack in the TCP request, but I'm not
> well versed enough in these methods to know.

TCP is just delivery mechanism, and has nothing to do with it.

--
William Park, Open Geometry Consulting, <opengeometry@yahoo.ca>
Linux solution/training/migration, Thin-client

Ashot Petrosian wrote:
> Is it possible to create a solution that creates dynamic thumbnails
> from iamges without downloading the entire image.
>
> This is on this newsgroup because its primarily a networking question.

It's not. - Read on.


> Can I request each image seperately and in the order that I want. In
> particular I know that sometimes when you download the image it goes
> from top to bottom, other times it goes from blurry to sharp. If I
> can request the order of information transfer, I could for each image
> just get the blurry versions and use that information to generate
> sharp small thumbs. These will not be ideal (ie average of
> surrounding pixels) but this will be fine. Thank you very much for
> any information.

Your question is about HTML and its techniques, this NG's primary
focus is on technical questions related to networking. HTTP, which
transports HTML, is in the application level, which nobody here
cares much about.

But to answer Your questions anyways: You can request each image
seperately. You will first ask for the HTML source of any given
page You want to view. H_T_ML is text only, and may contain refe-
rences to images that Your browser shall display within that text
area. Your browser (if not lynx) will recognize those references
and request each of those images in turn from the server. - So,
You need to get the page source first, parse it for image references
and then, ask for the images to be sent. That is what every browser
does.

When an image goes "from blurry to sharp", there are two concepts
for that: If it's ".gif", there's an "interlace" functionality in
this format (gif87a and gif89) which controls this. Non-interlace
will transfer the image line by line, while interlace will send
like $LINE_0; $LINE_N; $LINE_(N*2) .. $LINE_(N*n), and then start
from the top with $LINE_(0+1); $LINE_(N+1); $LINE_(N*2+1) and so on.
There's a similar thing for ".jpeg", but I'm no expert on that.
In HTML, there's also a "lowres" parameter for the <IMG> tag, where
You can specify a source that contains a smaller version (bytewise,
that is; that has nothing to do with the size of the displayed img)
of the to-be-displayed picture (like lower resolution and/or black-
/white) that will be loaded and displayed first, and only after all
such placeholders for the pictures of one page are loaded, the actual
images are being transmitted.

Only in this latter case, where "lowres" is used in an HTML <IMG>
tag, You can save bandwith by only downloading the lowres source;
with the interlace mechanism, You will have to load the entire file
first. - You can later resize it to save disk space, but to be able
to do so, You need to have the entire image.


> If this is possible is there an implementation, or do you guys have
> some tips as to how I could go about implementing this; I imagine it
> would be through some sort of hack in the TCP request, but I'm not
> well versed enough in these methods to know.

Again, You want to talk HTTP. That's applications. Go find an appro-
priate group for that. - Here, it's "TCP-County", with some more low-
level speakers. We don't understand Your high-level conversation...

Just kidding, Ashot, but I hope I got You on the right track.


Cheers, Jack.

--
----------------------------------------------------------------------
My personal reading of the string "MicroSoft" expands to "NanoWeak"...

Ashot Petrosian wrote:
> Is it possible to create a solution that creates dynamic thumbnails
> from iamges without downloading the entire image.
>
> This is on this newsgroup because its primarily a networking question.

It's not. - Read on.


> Can I request each image seperately and in the order that I want. In
> particular I know that sometimes when you download the image it goes
> from top to bottom, other times it goes from blurry to sharp. If I
> can request the order of information transfer, I could for each image
> just get the blurry versions and use that information to generate
> sharp small thumbs. These will not be ideal (ie average of
> surrounding pixels) but this will be fine. Thank you very much for
> any information.

Your question is about HTML and its techniques, this NG's primary
focus is on technical questions related to networking. HTTP, which
transports HTML, is in the application level, which nobody here
cares much about.

But to answer Your questions anyways: You can request each image
seperately. You will first ask for the HTML source of any given
page You want to view. H_T_ML is text only, and may contain refe-
rences to images that Your browser shall display within that text
area. Your browser (if not lynx) will recognize those references
and request each of those images in turn from the server. - So,
You need to get the page source first, parse it for image references
and then, ask for the images to be sent. That is what every browser
does.

When an image goes "from blurry to sharp", there are two concepts
for that: If it's ".gif", there's an "interlace" functionality in
this format (gif87a and gif89) which controls this. Non-interlace
will transfer the image line by line, while interlace will send
like $LINE_0; $LINE_N; $LINE_(N*2) .. $LINE_(N*n), and then start
from the top with $LINE_(0+1); $LINE_(N+1); $LINE_(N*2+1) and so on.
There's a similar thing for ".jpeg", but I'm no expert on that.
In HTML, there's also a "lowres" parameter for the <IMG> tag, where
You can specify a source that contains a smaller version (bytewise,
that is; that has nothing to do with the size of the displayed img)
of the to-be-displayed picture (like lower resolution and/or black-
/white) that will be loaded and displayed first, and only after all
such placeholders for the pictures of one page are loaded, the actual
images are being transmitted.

Only in this latter case, where "lowres" is used in an HTML <IMG>
tag, You can save bandwith by only downloading the lowres source;
with the interlace mechanism, You will have to load the entire file
first. - You can later resize it to save disk space, but to be able
to do so, You need to have the entire image.


> If this is possible is there an implementation, or do you guys have
> some tips as to how I could go about implementing this; I imagine it
> would be through some sort of hack in the TCP request, but I'm not
> well versed enough in these methods to know.

Again, You want to talk HTTP. That's applications. Go find an appro-
priate group for that. - Here, it's "TCP-County", with some more low-
level speakers. We don't understand Your high-level conversation...

Just kidding, Ashot, but I hope I got You on the right track.


Cheers, Jack.

--
----------------------------------------------------------------------
My personal reading of the string "MicroSoft" expands to "NanoWeak"...

jack wrote:
> Ashot Petrosian wrote:
> > Is it possible to create a solution that creates dynamic thumbnails
> > from iamges without downloading the entire image.

I really don't like to answer my own postings, but here's one more to
help You out:

As said, I'm not too familiar with JPEG, but what You can try is to
(1) find out about the technique JPEG uses for that "blurry to sharp"
transition You describe, (2) quantify the amount of data You want to
have of a given source, (3) create a script that will send an HTTP GET
request for that image (or, HTTP HEAD request which will return the
size of the file You're asking for, and only then GET it), and (4)
have this script RST'ing the connection after You got the amount of
data of that file that You need.

You will eventually end up with corrupted JPEGs, but that's not Your
interest, anyways.

And, as for how to break the HTTP transmission of that image, You will
need TCP/IP functionality all right. So this _is_ the correct NG for
this.

You knew all that before coming here, didn't You...? ;)


Cheers, JAck.

--
----------------------------------------------------------------------
My personal reading of the string "MicroSoft" expands to "NanoWeak"...