WO2002023401A2

WO2002023401A2 - A system and method for accessing web pages

Info

Publication number: WO2002023401A2
Application number: PCT/US2001/027647
Authority: WO
Inventors: Richard Hayton; David Halls
Original assignee: Citrix Systems, Inc.
Priority date: 2000-09-12
Filing date: 2001-09-07
Publication date: 2002-03-21
Also published as: AU2001288820A1; WO2002023401A3

Abstract

A method and system for reducing the bandwidth required in transmitting web pages to a browser is presented. The method includes a browser requesting from a proxy a first web page having a first content. A web page interface loads and transmits this first web page to the proxy and the proxy stores a copy of the first web page in its local memory. If the browser requests a second web page that is similar in content to the first web page, the proxy determines the differences between the content of the first web page and the content of the second web page after the proxy receives the second web page from the web page interface. The proxy then transmits these differences to the browser and the browser incorporates these differences with the first web page to create the second web page.

Description

A SYSTEM AND METHOD FOR ACCESSING WEB PAGES

FIELD OF THE INVENTION

The invention relates in general to accessing web pages and more specifically to a

system and method used to reduce the bandwidth required in transmitting web pages to a

browser.

BACKGROUND OF THE INVENTION

A user may request several web pages in sequence from a web browser. In such a

case, typically the browser requests the first web page from a server and the server loads

this first web page into memory from a persistent storage device. The server then

compresses the first web page before sending it back to the browser for decompression

and display. When the user selects a second web page, the browser usually discards the

first web page from its local memory and requests the second web page from the server.

The actions performed by the server and browser for the first web page are then repeated.

This method of accessing web pages occurs for each web page that the user selects

to

access, even if the web pages requested are similar. Therefore, the bandwidth between

the server and browser may be higher than required when transmitting a web page that is similar to the currently displayed web page. The present invention overcomes this waste

of bandwidth.

SUMMARY OF THE INVENTION

The invention features a system and a method that reduces the amount of data sent

over a network when a client computer accesses a web page that is similar to a

previously accessed web page. A user utilizes a browser to request from a proxy a first

web page having a first content. The first content includes a first web link that invokes a

request for a second web page having a second content. The proxy sends the request for

the first web page to a web page interface. The web page interface loads the first web

page into memory and transmits the first web page back to the proxy. The proxy ^'

determines using a predetermined criteria if the first content is sufficiently similar to the

second content of the second web page. If the proxy determines that the first content is

sufficiently similar to the second content, the proxy modifies the first web link to point to

a script routine. The proxy then stores the modified first content of the first web page in

its local memory for future use and sends the modified first web page to the browser.

If the user then utilizes the browseπtό'request the second web page, the browser

invokes the script routine. The 'script routine first transmits the request for the second

web page to the proxy. The proxy forwards this request to the web page interface and the web page interface returns the second web page having the second content to the

proxy. The proxy scans the second web page for web links that point to similar web

pages and modifies these web links to point to the script routine. The proxy then stores

the content of the modified second web page in its local memory for future comparisons.

The proxy then obtains the differences between the first and second web pages and

transmits the differences to the browser. The browser then uses the transmitted

differences and the content of the currently displayed first web page to produce

substantially a copy of the content of the second web page.

DESCRIPTION OF THE DRAWINGS i

The aspects of the invention presented above and many of the accompanying

advantages of the present invention will become better understood by referring to the

included drawings in which:

FIG. 1 is a block diagram of an embodiment of the system used to access two

similar web pages by transmitting the differences between the two web pages; and

FIGS. 2 A and 2B are sections of a flow diagram illustrating an embodiment of the

steps for accessing a second web page that is similar to a first web page. DESCRIPTION OF EMBODIMENTS

In brief overview and referring to Fig. 1, the network system, in one embodiment,

includes a server computer 50 (or server) in communication with a client computer 10

(or client). A user wishing to access a first web page performs an action on the client 10

to request the first web page from the server 50. For example, the user may use a

browser 20 to request the first web page. The server 50 loads the first web page into its

memory from a persistent storage device 60 and subsequently transmits the first web

page to the client 10. The client 10 then displays the first web page to the user.

If the user requests a second web page, the browser 20 of the client 10 similarly

requests the second web page from the server 50. In one embodiment the server 50 then

compares the first and second web pages and, if the second web page is similar to the

first web page, determines the differences between the two web pages. If the differences

between the first web page and -the second web page satisfy a predetermined criteria, as

described below, the server 50 compresses the differences between the two web pages

and transmits only the differences to the browser 20. The browser 20 then displays the

second web page on the client 10 by updating the first web page with the transmitted

differences. In more detail, the server 50 is in communication with a persistent storage device

60. The server 50 further includes a web page interface 40 in communication with a

proxy 30. The proxy 30 is in communication with the browser 20 over a communication

channel 15, and the browser 20 is accessed by the client 10.

In one embodiment, the client 10 uses the browser 20 to make a first request to

the proxy 30 for a first web page over the communication channel 15. The proxy 30 then

forwards the first request to the web page interface 40. The web page interface 40 loads

the first web page into the web page interface's 40 memory from the storage device 60

and provides the first web page.to the proxy 30. In order for the proxy 30 to identify

differences, the proxy 30 must receive the web page content in clear, that is, not

encrypted. Also, in another embodiment (shown in phantom), the proxy 30' is located on

another server 50' and is remotely located from the server 50 having the web page

interface 40. The proxy 30' communicates with the server 50 over a second

communication channel 45.

After the proxy 30 obtains the first web page, in one embodiment the proxy 30

modifies at least one reference in the first web page to another web page so that selecting

the modified reference calls a script routine. The script routine is software that the proxy

30 embeds within the first web page. Then the proxy 30 stores a copy of the modified

first web page in its local memory. In another embodiment, the proxy 30 stores the first web page in its unmodified state. The proxy 30 then sends the modified first web page,

which includes the embedded script routine, over the communication channel 15 to the

client 10. The client 10 then displays the first web page.

The client 10 then poses a second request to the server 50 for a second web page.

As in the first request, the web page interface 40 loads from the storage device 60 the

second web page corresponding to the second request and sends this second web page to

the proxy 30. The proxy 30 then determines the differences between the first web page

and the second web page. If the two web pages satisfy a predetermined criteria, the

proxy 30 compresses the differences and transmits the compressed differences to the

client 10 over the communication channel 15. As described in more detail below, the

client 10 decompresses the differences and displays a web page corresponding to the

second web page by incorporating the differences between the first web page and the

second web page into the previously displayed first web page.

Looking more closely at the steps performed by one embodiment and also

referring to FIGS. 2A and 2B, the user selects (step 200) a first web page PI that the user

wants displayed on the client 10 using the browser 20. The client 10 sends (step 205) a

request for the first web page PI to the proxy 30 and the proxy 30 forwards (step 210)

this request to the web page interface 40. The web page interface 40 loads (step 215) the first web page PI into its memory from the storage device 60. In another embodiment,

the web page interface 40 creates (step 215) the first web page PI.

Once the first web page PI is loaded, the web page interface 40 transmits (step

220) the first web page PI back to the proxy 30. The proxy 30 then modifies (step 225)

the first web page PI to enable difference comparisons between the first web page PI

and similar web pages. In modifying the web page, the proxy 30 initially scans the first

web page PI and searches for web links or other calls to other web pages (referred to

generally as web links) which, if selected, result in the first web page PI being replaced

by a second web page. For each of these web links, the proxy 30 determines if it is likely

that the web page referred to by the web link is similar to the first web page PI using a

heuristic program. The heuristic program uses a predetermined criteria to determine

whether two web pages are similar. Examples of the predetermined criteria include the

compressibility of the two web pages, the page names of the two web pages, and a meta

tag associated with the two web pages.

When using the compressibility criteria, the heuristic program computes the

differences between the two web pages. If the size of the differences between the two

web pages is substantially less t shan the size of the second web page, then the heuristic

program determines that the two web pages are similar. In another embodiment, the heuristic program uses the page names of the two web

pages as the predetermined criteria. The heuristic program compares the pathname of

the two web pages and considers similarly named web pages to be similar. Examples of

similar pathnames of two web pages include web pages in the same directory or web

pages generated by the same program executing on a web server (e.g., a servlet or Active

Server Page (ASP)).

In yet another embodiment, the heuristic program uses a meta tag criteria as the

predetermined criteria. Meta tags are a general mechanism for specifying attributes of

web pages and are typically used by web browsers 20 and readers of HTML source code.

For example, a meta tag can be added to a web page denoting whether a web page is

cacheable. A programmer can add meta tags to web pages manually or to the scripts that

generate the web pages. In this embodiment, the proxy 30 uses meta tags to denote a .

new attribute of web pages (e.g., the similarity between web pages). For example, a web

page has an added tag "<META isSimilarToLast>" which denotes similarity to the

previous web page. As another example, tags are added to sets of web pages, such as a

tag "<Meta name= OneOfSet' contents='ShoppingBasket'>". This meta tag includes

the keyword attribute OneOfSet' and the value 'ShoppingBasket' of the meta tag to

describe the web page. By using meta tags to denote similarity, a programmer or web page designer overrides the decision regarding similarity between two web pages

normally made by the proxy 30.

In a further embodiment, the proxy 30 determines similar web pages by keeping a

database of pairs of web pages found to be similar or different. For instance, a certain

meta tag, such as OneOfSet', is included within the web pages to indicate to the heuristic

program that the web pages are similar. In such a case, the proxy 30 maintains two

databases 48, 49, both of which are initially empty. The first database 48 includes

information on the previously loaded web pages that the proxy 30 loaded and the value

of the specific meta tag included within the web pages that can indicate to the heuristic

program that two web pages are similar (e.g., 'ShoppingBasket'). The second database

49 contains information relating two or more web pages (e.g., similar / dissimilar). In

the alternate embodiment described above, the remote proxy 30' determines similar web

pages by keeping a first database 48' and a second database 49'.

Specifically, if the heuristic program determines that an initial web page A has the

OneOfSet' meta tag, then the proxy 30 maps the initial web page A to the value of the

OneOfSet' meta tag (e.g., initial web page A -> ShoppingBasket). It should be noted

that the value of the meta tag may be a null value. If the initial web page A has a web

link to a reference web page B, the proxy 30 first consults the second database 49 to

determine if the proxy 30 has previously deemed the initial web page A and the reference web page B to be similar. If the second database 49 contains information indicating that

the initial web page A is similar to the reference web page B, then the proxy 30 modifies

the web link of the initial web page A referencing the reference web page B so that the

script routine is invoked when the browser 20 requests the reference web page B. If the

second database 49 contains information indicating that the initial web page A is

dissimilar to the reference web page B, then the proxy 30 does not consider the initial

web page A to be similar to the reference web page B. The proxy 30 does not modify

script routine is not invoked when the browser 20 requests the reference web page B.

If the second database 49 contains no information on the reference web page B, the

proxy 30 consults the first database 48. If the first database 48 has no information on the

reference web page B, the proxy 30 makes no decision regarding similarity between the

reference web page B and the initial web page A based on the meta tag heuristic and/or

the database entries. Instead, the proxy 30 employs one of the other previously described

heuristics (e.g., compressibility and/or page names) to determine whether the initial web

page A is similar to the reference web page B.

If the first database 48 contains the, same value of the meta tag for the reference

web page B that is associated with the initial web page A (e.g., 'ShoppingBasket') and

the values are not equal to null, then the proxy 30 deems the initial web page A similar to - li ¬

the reference web page B. Two web pages having meta tag values that are equal to null

are not considered similar by the proxy 30 in order to ensure that only specific meta tag

values are considered equivalent. In another embodiment, the proxy 30 considers web

pages similar when each web page has a meta tag value equal to null. The proxy 30 then

modifies the web link of the initial web page A referencing the web page B so that the

script routine is invoked when the browser 20 requests the reference web page B.

If the first database 48 contains different values of the meta tags associated with

the initial web page A and the reference web page B, then the proxy 30 does not consider

the initial web page A to be similar to the reference web page B. Therefore, the link of

the initial web page A referencing the reference web page B is not modified. It should

be noted that a modified initial web page A can have some modified web links to web

pages and some unmodified web links to other web pages. Besides traditional databases,

the proxy 30 can alternatively use memory data structures or files stored on a local disk

to keep the first database 48 and the second database 49. Although the proxy 30 employs

a first database 48 and a second database 49 to maintain the previously described

information on the web pages, the proxy 30 can alternatively use a single database or

multiple databases to store the meta tag information and the similarity information.

To increase efficiency, the heuristic program can be optimistic; that is, the

heuristic program on the proxy 30 assumes that a web link results in a similar web page. For example, if the heuristic program uses the page name criteria, the heuristic program

can assume that any web pages within the same directory are similar. If the assumption

made by the heuristic program is incorrect (i.e., the two web pages are not similar), the

browser 20 still displays the correct second web page because the proxy 30 in this

situation (incorrect non-similarity determination) transmits the second web page to the

client 10.

During operation of a further embodiment, the heuristic program employs the page

name criteria to guess whether the two web pages are similar. If the proxy 30 has

previously guessed that the two web pages are similar using the page name criteria and

then follows the web link to the second web page, the proxy 30 retrieves the second web

page from the web page interface 40 and applies one or a combination of the other

criteria (e.g., compressibility and/or meta tag criteria) to determine whether the proxy 30

should transmit the second web, page or the differences between the two web pages. As

described in more detail below and in a further embodiment, the proxy 30 updates the

second database 49 when the proxy 30 makes its final decision on whether to transmit

the second web page or the differences between the two web pages. To check that the

heuristic was helpful in determining similarity, the proxy 30 can employ the

compressibility criteria even if no meta tags exist in the web pages. In contrast, if the proxy 30 follows a web link to a second web page and the proxy

30 has previously determined that the two web pages are not similar, then the proxy 30

retrieves the second web page from the web page interface 40 but does not compare the

two web pages. The proxy 30 at this point can examine the second web page to

determine if the second web page contains a meta tag indicating similarity with the first

web page. If such a meta tag is found, the proxy 30 can store this information in the

previously described first database 48 for future comparisons.

If the proxy 30 uses the heuristic program and determines that a web link refers to

a web page similar to the first web page Pi, the proxy 30 modifies (step 225) the first

web page PI so that the activation of that web link within the first web page PI calls a

script routine that executes on the browser 20. When the user clicks on the web link, the

browser 20 invokes the script routine.

In one embodiment, the script routine is software written in JavaScript, a scripting

language used to develop client-side Internet applications. It should be understood by

those skilled in the art that the script routine can be written in any computer language so

long as the browser 20 can interpret and execute the script. An example of a script

routine is to replace a reference "<a href="foo">click here</a>" with "<span

onClick="goGetIt('foo')">click here</span>." In this example, goGetltQ is a JavaScript

function added to the first web page PI which sends the second web page P2 request to the proxy 30. The proxy 30 responds with either the second web page P2 or the

differences between the first web page PI and the second web page P2. The JavaScript

function then performs additional processing, as described below, to recreate the second

web page P2 before the browser 20 displays it. Other references, such as form submits

(the button or method used by the user of the browser 20 to submit a form to the server

50) can be treated in the same way.

As a more specific example with a form submit, a Submit button (used for

searches, etc), may call the script routine when the user invokes the function. An

example of a script routine for a Submit button is to replace the JavaScript line "<input

id=GoBtn type=submit value="Go">" with "<input id=GoBtn type=button

value="Go"onClick="goGetForm()">." In this example, goGetForm() is a JavaScript

function provided in the script routine to call the proxy 30. Furthermore, if the software

code for activating other browser 20 functions (e.g., a Refresh button) is accessible to the

proxy 30, then the proxy 3Q can modify these web page buttons as described above.

Because the proxy 30 will need the contents of the first web page PI later, the

proxy 30 then stores (step 230) a copy of the modified first web page PI in its local

memory. By storing the first web page PI after modification, future comparisons of the

first web page PI with a second web page are more accurate because the proxy 30 does

not deem the same modifications on both web pages as differences. Additionally, the proxy 30 marks its copy of the first web page PI to indicate to which client 10 the proxy

30 sent the first web page PI. This is necessary in a system with more than one client 10

requesting the same web page PI .

The proxy 30 then sends (step 235) the first web page PI to the browser 20 over

the communication channel 15 and the browser 20 displays (step 240) the first web page

PI on the client 10. In another embodiment the proxy 30 compresses the first web page

PI prior to transmitting it to the browser 20. If the user then selects (step 245) a second

web page P2 from a web link that had been modified by the proxy 30, the selection

invokes the script routine (phantom 250) on the client 10. Once invoked, the script

routine transmits (step 255) the second request to the proxy 30. The second request for

the second web page P2 transmitted by the script routine is a different request than the

first request for the first web page PI . For example, a first request transmitted by the

browser 20 is "HTTP GET /some/page." For the second request, the script routine uses a

special name (e.g., "special name") to invoke a servlet or other software to calculate the

differences on the proxy 30. An example of a second request transmitted by the script

routine is "HTTP GET /special name? 'some/page'."

The script routine also notifies the proxy 30 to compare the currently displayed

first web page PI, which the proxy 30 indexed, with the requested second web page P2

by including the "special name" in the second request. The script routine also notifies the browser 20 to open a non-displayed window in which the differences between the

first web page PI and second web page P2 are stored. In this way, the displayed first

web page PI is left intact. The browser 20 can then recreate the second web page P2

from the transmitted differences stored in the non-displayed window and the displayed

first web page PI.

The proxy 30 again forwards (step 260) the request (e.g., the second request for the

second web page P2) to the web page interface 40. The web page interface 40 creates or

loads (step 265) the second web page P2 and transmits (step 270) the second web page

P2 back to the proxy 30. The proxy 30 next modifies (step 275) the web links in the

second web page P2 to invoke the script routine (phantom 250) using the same heuristic

program the proxy 30 used to modify the web links in the first web page PI . The proxy

30 then stores (step 280) the modified second web page P2 and deletes the previously

stored web page. As previously described with respect to the first web page PI, by

storing the second web page P2 after modification, future comparisons of the second web

page P2 with another web page are more accurate because the proxy 30 does not deem

the same modifications on both, web pages as differences. In another embodiment, the

proxy 30 modifies (step 275) the second web age P2 after storing (step 280) the second

web page P2 in its local memory. In one embodiment, the proxy 30 calculates the differences between the first web

page PI and the second web page P2 by treating the contents of both web pages as

sequences of characters and comparing the two pages on a character by character basis.

In another embodiment, the proxy 30 considers the contents of the two web pages as

trees of HTML elements. Examples of HTML elements are web links and characters. A

few specific examples of HTML elements are "<TEXT background=red>hello

world</TEXT> and <LIST>[child tags of type <LI>] </LIST>." When data is organized

in a tree-like structure, each element in a tree is referred to as a node. A parent node is a

node that has one or more children nodes. Nodes that have no children are called leaf

nodes. In this embodiment, the proxy 30 compares the trees for common leaves and

nodes to obtain the differences between the web pages.

The proxy 30 then compresses (step 285) the differences between the first web

page PI and the second web page P2 using compression software. The proxy 30

subsequently determines if the transmittal of the differences between the first web page

PI and the second web page P2 to the browser 20 is less wasteful of bandwidth than the

transmittal of the second web page P2 itself. To help in this determination, the proxy 30

compresses the second web page P2. The proxy 30 then compares the size of the

compressed differences to the size of the compressed second web page P2. If the proxy

30 concludes that the compressed differences are not smaller than the compressed second web page P2, then the proxy 30 sends the compressed second web page P2 to the client

10. ^;

As briefly discussed above, in another embodiment the proxy 30 updates the

second database 49 when the proxy 30 makes a final decision as to whether the

differences between the two web pages Or the content of the second web page P2 is sent

to the client 10. More specifically, the proxy 30 updates (step 285) the second database

49 each time a difference between the first web page PI and the second web page P2 is

calculated. If the heuristic program initially determines that the first web page PI is

similar to the second web page P2 and the proxy 30 transmits the differences between the

first web page PI and the second web page P2 to the client 10, the proxy 30 denotes in

the second database 49 that the web pages are similar (e.g., first web page PI, second

web page P2 -> similar). If the heuristic program initially determines that the first web

page PI is similar to the second web page P2 and then the proxy 30 determines that the

two web pages are actually dissimilar (and therefore transmits the second web page P2 to

the client 10 rather than the differences), then the proxy 30 denotes in the second

database 49 that the web pages are dissimilar (e.g., first web page PI, second web page

P2 -> dissimilar).

If the heuristic program initially determines that the first web page PI is not

I) similar to the second web page P2, then the proxy 30 does not compute the differences between the first web page PI and the second web page P2 and therefore does not update

the second database 49. In another embodiment, the proxy 30 computes the differences

between the first web page PI and the second web page P2 to update the second database

49 and thereby improve future similarity determinations by the heuristic program.

Because the heuristic program uses the first and second database 49 to check the

heuristic program's determination of similarity between web pages, the heuristic

program can be optimistic; that is, the heuristic program on the proxy 30 assumes that a

web link results in a similar web page. Ho,wever, the heuristic program still follows the

similarity decisions in the second database 49 that is updated after assuming that a web

link results in a similar web page.

Otherwise, the proxy 30 sends (step 295) the compressed differences between the

two web pages to the client 10. The proxy 30 also discards (step 290) the stored copy of

the first web page PI. In another embodiment, the proxy 30 sends the compressed

differences between the two web pages to the client 10 if the proxy 30 concludes that the

compressed differences are smaller than the compressed second web page P2 by a

predetermined threshold, such as by a predetermined number of bytes. In another

embodiment, the proxy 30 does not compress the second web page P2 and therefore does

not compare the compressed differences to the compressed second web page P2.

Instead, the proxy 30 always transmits the compressed differences to the client 10. i While the proxy 30 is implementing step 260 through step 295, the script routine

executing on the client 10 awaits a response from the server 50. Once the browser 20

receives the data from the proxy 30, the browser 20 decompresses the compressed data

using decompression software. In the case where the browser 20 invoked the script

routine and therefore the proxy 30 transmitted the differences, the browser 20 recreates

(step 297) the second web page P2 by incorporating the differences between the first web

page PI and the second web page P2 into the previously displayed first web page PI.

I

In another embodiment, the first web page PI is capable of modifying itself with

an embedded modifying script function and so the content of the first web page PI is

capable of changing often. Therefore, because the content of the first web page PI

changes, the browser 20 stores an original, copy of the first web page PI to allow a

comparison for the differences between the original first web page PI and the second

web pages. In an embodiment in which the proxy 30 does not compute the differences

between this first web page PI and the second web page P2, the browser 20 does not

store an original copy of the first web page PI because the proxy 30 transmits the second

web page P2 to the browser 20.

In one embodiment described above in which the proxy 30 calculates the

differences between the first web page PI and the second web page P2, the proxy 30

treats the contents of both web pages as sequences of characters and compares the two pages on a character by character basis. Therefore, the differences are sent in the form of

textual modifications. For example, the proxy 30 performs a Unix "diff ' command to

obtain the differences between the two web pages. More specifically, the transmitted

differences instruct the browser 20 to insert XXXX at position Y and delete AAA

characters at position ZZZ, where X and A represent characters and Y and Z represent

positions on the first web page PI or second web page P2. Upon receiving these

transmitted differences, the browser 20 performs these modifications on the previously

displayed first web page PI to create a new second web page P2. The browser 20 uses a

standard JavaScript function "d^'ocument.setInnerHTML (<html source>)" to redisplay

the new second web page P2. The browser 20 then discards (step 298) the unneeded first

web page PI from its local memory and displays (step 299) the new second web page P2

on the client 10.

In another previously described embodiment, the proxy 30 considers the contents

of the two web pages as trees of HTML elements. Therefore, the differences are sent in

the form of structured differences. The browser 20 modifies the displayed first web page

PI without removing the displayed first web page PI from the display of the client 10.

It will be appreciated that the embodiments described above are merely examples

of - 11 - -

the invention and that other embodiments incorporating variations therein are considered

to fall within the scope of the invention. In view of the foregoing, what is claimed is:

Claims

1. A method for accessing web pages comprising the steps of:

receiving a request for a first web page having a first content;

receiving a request for a second web page having a second content; and

transmitting differences between said first content of said first web page and said

second content of said second web page to a browser.

2. The method of claim 1 further comprising the steps of:

storing said first web page;

modifying said first content of said first web page to create a modified first web

!

page having a modified first content; and

obtaining said differences between said first content and said second content.

3. The method of claim 2 wherein said first content of said first web page includes a

first web link that invokes said request for said second web page.

4. The method of claim 2 further comprising determining with a first predetermined

criteria if said first content is sufficiently similar to said second content.

5. The method of claim 4 wherein said modifying said first web page further

comprises modifying said first content by changing said first web link to point to a

script routine when said first predetermined criteria is satisfied.

6. The method of claim 5 wherein said script routine is a program that causes said

proxy to obtain said differences between said first content and said second content

when said script routine is executed.

7. The method of claim 4 wherein said first predetermined criteria is amount of

compressibility of said first web page and said second web page.

8. The method of claim 4 wherein said first predetermined criteria is similarity in

pathnames of said first web page and said second web page.

9. The method of claim 4 wherein said first predetermined criteria is similarity in a

tag associated with said first web page and said second web page.

10. The method of claim 4 wherein said first predetermined criteria is a database.

11. The method of claim 10 wherein said database further comprises information

denoting similarity between said first web page and said second web page.

12. The method of claim 2 wherein said obtaining said differences further comprises

comparing each character of said first content with each respective character of

said second content.

13. The method of claim 2 wherein said obtaining said differences further comprises

comparing an element of said first content with an element of said second content.

14.The method of claim 2 wherein said obtaining said differences further comprises maintaining a database denoting similarity between said first web page and said

second web page.

1 15. The method of claim 1 further comprising the step of compressing said differences

between said first content and said second content.

16. The method of claim 15 further comprising the step of discarding said first web

page after compressing said differences.

17. The method of claim 1 wherein said step of transmitting occurs when said

differences are less than a second predetermined criteria.

18. The method of claim 1 fu !rther comprising the step of maintaining a record

indicating that said first web page was transmitted to said browser.

19. A method for accessing web pages comprising the steps of:

displaying a first web page having a first content;

transmitting a request to obtain a second web page having a second content;

receiving differences in content between said first and second web pages in

response to said request; and .

incorporating said differences into said first content of said first web page to

produce substantially a copy of said second content of said second web page.

20. The method of claim 19 further comprising communicating with a proxy to obtain said differences between said first content of said first web page and said second content

of said second web page.

21. The method of claim 19 further comprising the step of decompressing said

differences.

22. The method of claim 19 further comprising the step of storing said first web page

until said copy of said second web page is produced.

I 23. A system for accessing web pages comprising:

a browser displaying a first web page having a first content and transmitting a

request for a second web page leaving a second content; and

a proxy in communication with said browser to receive said request for said

second web page and to transmit differences between said first content and said second

content to said browser in response to said request.

24. The system of claim 23 further comprising a web page interface in communication

with said proxy and storage, said web page interface accessing said first and

second web pages from said storage and said web page interface transmitting said

first and second web pages to said proxy.

1

25. The system of claim 23 further comprising a compressor compressing said

differences between said first content and said second content.

26. The system of claim 23 wherein said browser comprises a decompressor

decompressing said differences between said first content and said second content.

27. The system of claim 23 further comprising a comparator comparing each character

of said first web page with each respective character of said second web page to

obtain said differences.

28. The system of claim 23 further comprising a comparator comparing each element

of said first content with e 1 ach element of said second content to obtain said

differences.

29. The system of claim 23 further comprising a storage storing said first web page for

future comparisons with said second web page and storing an indicia with said

first web page indicating to said browser that said proxy transmits said first web

page.