Ex Parte Sundaresan

Board of Patent Appeals and Interferences

Aug 27, 2009

09672304 (B.P.A.I. Aug. 27, 2009)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________
Ex parte NEELAKANTAN SUNDARESAN
____________
Appeal 2008-005447
Application 09/672,3041
Technology Center 2100
____________
Decided: August 27, 2009
____________
Before LEE E. BARRETT, JOHN A. JEFFERY, and STEPHEN C. SIU,
Administrative Patent Judges.

BARRETT, Administrative Patent Judge.

DECISION ON APPEAL
This is a decision on appeal under 35 U.S.C. § 134(a) from the final
rejection of claims 1-23. We have jurisdiction pursuant to 35 U.S.C. § 6(b).
We reverse.

1 Filed September 29, 2000, titled "Method and System for
Selectively Accessing Files Accessible Through a Network." The real party
in interest is International Business Machines Corporation. Supplemental
Brief filed November 6, 2007 (Br.), 3.
Appeal 2008-005447
Application 09/672,304

2
STATEMENT OF THE CASE
The invention
A conventional Internet search engine accesses websites on the
network. The search engine downloads data from the website and archives
selected downloaded data. One can use the search engine to search for a
particular website containing desirable information by entering a query into
the search engine. The search engine will search its archived data and return
websites in its archived database which relate to the query. Spec. 3.
The dynamic nature of the Internet results in websites being updated
regularly. Consequently, data which was on the website when the search
engine initially visited the website may no longer be there. Alternatively,
the data may be outdated. Further, the website may no longer exist or its
URL may have changed. As a result, data archived by the search engine
could become invalid. In order for the search engine to be a useful tool, the
search engine must periodically update its archived data. Spec. 3.
A conventional search engine uses a web crawler (e.g., a "robot",
"spider", "ant", etc.) to visit (i.e., access) a server on a network. The spider
"crawls" from a homepage (i.e., the first or main webpage) of a website to
the various subpages linked from the homepage. As the web crawler visits
the various homepages with subpages, data on the pages are selectively
archived by the search engine. Spec. 3.
The typical crawlers visit web sites at regular intervals, for example,
every 30 days. If a web crawler accesses a website which has not been
updated since the last time the web crawler visited, the web crawler would
Appeal 2008-005447
Application 09/672,304

3
presume that the data previously archived is still valid. This may be
erroneous. That is, one disadvantage with current web crawler technology is
that the web crawler does not know when a website is scheduled to be
updated. Depending on how often a website is updated, the web crawler's
archived data could be very outdated by the time the web crawler returns.
On the other hand, frequent web crawler visits to websites not frequently
updated consume valuable computer resources. Spec. 3-4.
The invention provides a method and system for determining when
and how often a web crawler should return to a website. Websites often
contain Channel Definition Format (CDF) files which include data used by
websites to specify how often and what parts of the site will be "pushed"
(e.g. e-mailed) directly to a registered subscriber. A typical CDF file is an
Extended mark-up language (XML) file containing various elements
referred to as tags. The CHANNEL tag specifies the Universal Resource
Locator (URL) on the website that corresponds to that CHANNEL. The
SCHEDULE tag indicates when a channel should be updated. A channel
may have a subchannel with information about when the page was last
modified. Spec. 1-2.
The invention downloads the CDF file (the claimed "first file"),
storing the CHANNEL (indicating the "second file") and SCHEDULE
information. Next the invention uses the SCHEDULE information in the
CHANNEL tag to decide when to visit the website next. Spec. 8.

Appeal 2008-005447
Application 09/672,304

4
The claims
Claim 1 is reproduced below:
1. A method for searching files stored on a network, comprising:
downloading a first file on the network from a server to a client;
accessing time data from within the first file; and
setting an accessing time to access a second file on said server
based on said time data from the first file, wherein said time data
includes an actual time when said second file is scheduled to be
updated.

The references

Yamane 6,167,436 Dec. 26, 2000
(filed Aug. 5, 1997)
Ronning US 2003/0212992 A1 Nov. 13, 2003
(effective filing date Jan. 27, 2000)

The rejections
Claims 1-23 stand rejected under 35 U.S.C. § 103(a) as unpatentable
over Ronning and Yamane.
The rejection of claims 1-23 under 35 U.S.C. § 112 ¶ 1 as failing to
comply with the written description requirement, although stated at page 3
of the Examiner's Answer, is apparently withdrawn as noted at page 14.

Appeal 2008-005447
Application 09/672,304

5
ISSUES
Appellant argues that the references do not teach or suggest the
following limitations:
(1) "accessing time data from within the first file; and setting an accessing
time to access a second file on said server based on said time data
from the first file, wherein said time data includes an actual time when
said second file is scheduled to be updated" (claims 1 and 17);

(2) "accessing time data from within said first file; and setting an
accessing time to re-access the server based on said time data from the
first file, wherein said time data includes an actual time when a
second file is scheduled to be updated" (claim 7);

(3) "means for accessing time data from within the first file; and means
for setting an accessing time to access a second file on said server
based on said time data from the first file, wherein said time data
includes an actual time when said second file is scheduled to be
updated" (claim 23).

The limitations in claim 1 are representative of the rest of the claims.
Accordingly, the issue is:
Has Appellant shown that the Examiner erred in concluding that the
references teach or suggest "accessing time data from within the first file;
and setting an accessing time to access a second file on said server based on
said time data from the first file, wherein said time data includes an actual
time when said second file is scheduled to be updated," as recited in
representative claim 1?

Appeal 2008-005447
Application 09/672,304

6
PRINCIPLE OF LAW
Obviousness requires that the differences between the subject sought
to be patented and the prior art are such that the subject matter as a whole
would have been obvious to one of ordinary skill in the art. 35 U.S.C.
§ 103(a).

FACTS
Ronning
Ronning describe an agent software application for controlling
distribution of files and managing updates to files. Abstract.
Figure 15 of Ronning shows a general settings screen 738. "Selection
of a section 740 permits the user to instruct the agent to search for updates
on a periodic basis. In particular, the user may enter a number of days in
section 741 by which they want the agent to automatically search for update
to the files on the user's machine . . . ." ¶ [0072].

Yamane
Yamane describes that network robots acquire data items dispersed
over a network by referring to links. Col.1, ll. 11-13.
A problem with conventional network robots is that even where the
contents of the data have not been updated, they may be still acquired,
resulting in the execution of useless data acquisition. Col. 1, ll. 58-61.
One object of Yamane is to suppress unnecessary acquisition of
un-updated data items. Col. 2, ll. 16-20.
Appeal 2008-005447
Application 09/672,304

7
Yamane uses an update time from an acquired file and an update
history to predict the next update time so that the system will not acquire
data before it is updated. "A first process is to extract the update time from
the acquired data, a second process is to determine whether or not the
acquired data has been updated, and a third process is to calculate the next
update prediction time from the update history." Col. 6, ll. 36-40.
The first process is to determine whether information on the latest
update time exists in the transferred data. Col. 6, ll. 48-50. The acquisition
time, latest update time, and hash value from multiplying the acquired data
by a hash function are stored in the update history table. Col. 7, ll. 10-15.
When information on the latest update time does not exist, the second
process of determining whether the acquired data has been updated may be
done by comparing a hash value of the acquired data to a hash value of the
latest update time. Col. 6, l. 56 to col. 7, l. 9. The values are stored in the
update history table. Col. 7, ll. 10-15.
The third process of calculating the next update predication time T
from the update history involves calculating a "prediction update interval"
by subtracting the next latest update time from the latest update time
registered in the history table, and then calculating the "next update
prediction time T" by adding the prediction update interval to the latest
update time. Col. 8, ll. 1-22.

Appeal 2008-005447
Application 09/672,304

8
CONTENTIONS
The Examiner finds that Ronning describes accessing a first file on a
network at paragraph [0035], and describes setting an access time to access a
second file on a server at Figure 15, elements 740-743. Final Rej. 7. The
Examiner finds that Ronning does not teach accessing time data from within
the first file and the time data being an actual time for future updates. Id.
The Examiner finds that "Yamane[] teaches a system where it predicts a
future update including setting an actual time and date of when this update
will occur (col. 4, line 63 - col. 5, line 3; Col. 5, Lines 48-58)." Ans. 8. The
Examiner concludes that it would have been obvious to modify Ronning
with the teachings of Yamane for the next prediction update to reduce traffic
congestion by minimizing the number of times the agent has to search the
server for updates. Id.
Appellant argues that all of the independent claims recite an actual
time when a second file is scheduled to be updated and that this limitation is
not taught or suggested in the references. Br. 11.
Appellant argues that the files downloaded by Yamane only include
data on when a file was last updated and do not include any schedule at all.
Reply Br. 5. It is noted that Yamane discloses trying to predict when files
might be updated based only on data which is downloaded from the servers
based upon time when these files have been updated in the past. Id. at 5-6.

Appeal 2008-005447
Application 09/672,304

9
ANALYSIS
We first deal with the Examiner's finding regarding Ronning.
The Examiner finds that Ronning teaches "setting an accessing time to
access a second file on a server." Ans. 7. However, the full limitation
requires "setting an accessing time to access a second file on said server
based on said time data from the first file." Ronning sets the accessing time
based on user input, not based on time data from the first file, so the
Examiner's statement is not persuasive as to the actual claim limitation.
The Examiner also states that Ronning teaches that the user can set
parameters related to control of file updates, including scheduling of
downloads, and thus, "[t]echnically, Ronning along [sic, alone] teaches the
limitation 'setting an accessing time to access a second file on a server based
on time data that includes an actual time when the second file is scheduled to
be updated' as claimed." Id. at 8. However, the Examiner fails to address
that the "time data" is time data from the first file, "wherein said time data
includes an actual time when said second file is scheduled to be updated."
We agree with Appellant that, "[t]he system that is disclosed by the Ronning
et al. reference does not teach or suggest downloading a file from a server
which includes a scheduled update time." Reply Br. 8. Therefore, Ronning
is not of any use to the disputed limitations.
Next we look at Yamane.
Appellant argues that "like the Ronning et al. reference, the Yamane
et al. reference does not teach or suggest downloading a first file from a
server that includes an actual time when the second file is scheduled to be
Appeal 2008-005447
Application 09/672,304

10
updated." Reply Br. 8. We agree. Yamane describes that the time data
within the first file is the "latest update time," col. 6, ll. 35-47, and is not "an
actual time when said second file is scheduled to be updated" as claimed.
Yamane describes that the "latest update time" is used together with
previously stored "latest update times" to calculate a "prediction update
interval" which is added to the latest update time to calculate a "next update
prediction time T." As noted by Appellant, the update time in Yamane is
generated locally, not downloaded. Reply Br. 9. Manifestly, the claimed
invention does not require "predicting" or "calculating" a next update time
because "an actual time when said second file is scheduled to be updated" is
part of the time data from the first file.
The point of contention in the rejection appears to be the Examiner's
interpretation that the calculated predicted "time when said second file is
scheduled to be updated" in Yamane meets the claimed "actual time when
said second file is scheduled to be updated" limitation. We understand that a
"predicted time" is an "actual time" in the sense that it is real. However,
considering the limitation as a whole, it is clear that a "predicted time when
said second file is scheduled to be updated" is not the same as an "actual
time when said second file is scheduled to be updated." There is a
difference between predicting that a file will be updated in one week and
knowing for certain that the file is scheduled to be updated in one week: a
predicted time may not, and probably will not be the actual time.
Moreover, the Examiner's interpretation does not account for the fact
that the limitation of "setting an accessing time to access a second file on
Appeal 2008-005447
Application 09/672,304

11
said server based on said time data from the first file, wherein said time data
includes an actual time when said second file is scheduled to be updated"
requires that "said time data [from the first file] includes actual time when
said second file is scheduled to be updated." We have noted that the time
data in Yamane is the "latest update time" and not "an actual time when said
second file is scheduled to be updated" as claimed. These points have been
extensively argued by Appellant in the main Brief and the Reply Brief.
We are persuaded by Appellant's arguments that the Examiner erred.

CONCLUSION
We are persuaded by Appellant's arguments that the Examiner erred in
concluding that the references teach or suggest "accessing time data from
within the first file; and setting an accessing time to access a second file on
said server based on said time data from the first file, wherein said time data
includes an actual time when said second file is scheduled to be updated," as
recited in representative claim 1. The rejection of claims 1-23 is reversed.

REVERSED

rwk

MCGINN INTELLECTUAL PROPERTY LAW GROUP, PLLC
8321 OLD COURTHOUSE ROAD
SUITE 200
VIENNA, VA 22182-3817