Does the Abstract matter any more?

The PCT rules and Applicant’s Guide talk about what should go into an Abstract.  In the PCT system it is up to the International Searching Authority to enforce the rules about Abstracts.  WIPO is inviting comments (see Circular PCT 1517) on whether the rules and guide should be revised.  Such revisions, if they were to happen, would change what the ISA looks for when it reviews an Abstract.

Comments are due September 30, 2017.  If you would like to participate in a listserv committee to prepare and file comments, join the PCT listserv (if you are not already a member) and join in the discussion there.  What follows are a few thoughts about Abstracts.

In the old days the Abstract certainly did matter a lot.  When I was first in practice, I used to hop the Amtrak Metroliner from Manhattan to come down to the Patent Office.  I would go to the USPTO search room in Crystal Plaza 3-4 and I would flip through “shoes”.  Each “shoe” was a big stack of pages.  Each page in a shoe was the front page of a patent.  The patents in a given shoe were in that shoe because they were all in the same patent class and subclass.

The searcher (such as myself) would flip through the shoes, and for each page, the searcher would glance at the Abstract and the published figure to decide whether a particular patent deserved further scrutiny.  In those days the Abstract really really mattered a lot.  A poorly worded Abstract would fail to summarize the novel disclosure in the patent.  A searcher could miss a relevant patent for no better reason than that the Abstract had been poorly (or cleverly, if the filer wished to avoid attention) drafted.  A searcher could likewise miss a relevant patent for no better reason than that the Figure selected to go on the front page might be poorly chosen.

Three important things have changed since those old days.

First, the Metroliner is gone, replaced by the Acela.  The last Metroliner ran on October 27, 2006.  The Metroliner managed the trip from New York to Washington in as little as 2½ hours.  As a bit of backwards progress, the fastest scheduled Acela service takes 2¾ hours to get from New York to Washington.  Sigh.

Second, Crystal Plaza 3 and 4 are gone!  Well, not gone exactly, but I was astonished to learn that on June 26, 2004 they got renamed as 2100 Crystal Drive and 2200 Crystal Drive.   The timing of this name change seems odd to me.  The new public search room in the Alexandria patent office did not open until September 27, 2004.  So the USPTO must still have been in the old Crystal City search room during this odd time from June 26, 2004 to September 27, 2004.  During those last three months of operation of the old search room, I guess searchers went to a building where the big brass “3” and “4” had already been taken down and new “2100” and “2200” brass numerals must have been in place.

Nowadays you can still go to Crystal City but you can’t go and stand someplace and see a building with “Crystal Plaza 3” or “Crystal Plaza 4” written on it.  And as for the underground walkways of Crystal City that we all knew and loved, the underground walkways are still there, but the maps in the walkways no longer show a “Crystal Plaza 3” or “Crystal Plaza 4”.  It’s like a bit of history got erased.

Who besides me recalls the days of the old public patent search room in Crystal City 3-4?  Post a comment below.

Third, computers happened and the Internet happened.  Nowadays, if you are doing a patent search, it does not matter at all how the Abstract was drafted or which figure happens to have been selected to go on the front page.  The search is based upon key words or patent classes or the assignee.  And the key words might be in the claims, or might be in the specification.  The searcher gets to pick.  The search engines also permit the searcher to specify that the key words are in the Abstract, but nobody does that.  The Abstract is completely irrelevant to searching these days.  The Abstract could be 150 words or 350 words or 50 words, and nothing about this would make any difference to the results of a modern-day patent search.  It could be well worded or poorly worded and nothing about it would make any difference for a search.

Which brings us to this Circular.  I am only guessing, but what I’d guess is that what’s going on here is two things:

  • The Searching Authorities are tired of having to carry out the duty (imposed by the searching authority guidelines, see pages 150-152) of enforcing the word count limitation, and the duty (also imposed by the searching authority guidelines) of having to review the Abstract to decide whether it is necessary to revise or redraft the Abstract.  They know that nobody actually uses the Abstract any more for searching.  They wish that they could be relieved of having to spend time tinkering with the Abstract.
  • The IB is tired of having to carry out the duty (imposed by the IB guidelines) of translating the text in the front-page figure.  It costs money to do this translating but worse, it can’t be completely computerized because someone has to go in with an image editor and erase the first-language text and paste in the second-language text.  The IB knows that nobody actually relies upon the front-page figure any more for searching.  They wish that filers could be told to stop putting any text at all into the front-page figure, but only reference numerals.

As for the word-count limitation in the Abstract, the guilty parties are the filers who have not yet migrated to the software for generating Requests.  The software counts the words in the Abstract and scolds the filer if the word count has not been complied with.  Even now something like 30% of US filers fail to use the software.  They stubbornly use a fillable PDF form instead, and nothing about the fillable form workflow enforces the word count.  Those are the filers who violate the word-count limitation.

As for the text in the publication figure, the software for generating Requests requires the filer to provide, in computer-readable form, the text in that figure.  (See “framus” and “clevice” at right.)  This text then goes to the IB’s translation contractors who then carry out a machine translation.  Still a human operator must carry out image editing steps to erase the old text and insert the new (machine translated) text.  Again there are those stubborn 30% of US filers that fail to use the software.  Those filers force the IB to manually scrutinize the front-page figure and to hand-key its text into the machine translation system.

Some day the stubborn 30% of US filers that fail to use ePCT (or PCT-SAFE) to file will migrate to ePCT.  When this happens, there will no longer be word-count violations in the Abstracts.  And when this happens, consistently the filer will have provided the text from the front-page figure in computer-readable form for machine translation.

This migration would still leave unchanged two of the presenting problems.  The ISAs would still wish they would not have to spend time reviewing Abstracts to figure out whether there is a need to revise or replace the Abstracts.  And the IB would prefer not to have to pay humans to carry out image editing of the front-page figures.

The Applicant’s Guide and the search guidelines could be revised to reduce or eliminate a requirement that the ISA review the Abstract for its substantive content.

The rules and Applicant’s Guide could be revised to more strongly discourage the use of text in the figures.

In my own practice I try very hard not to have any text in the figures, for the simple reason that keeping text out of the figures reduces translation costs down the line.  In today’s world of increasing internationalization of patent filing, it would probably actually be a money-saver for filers if everyone were to eliminate all text from all figures and use reference numerals only.

But it is natural to imagine that anybody would prefer when they are reviewing a patent, that the figures contain text rather than mere reference numerals.  A picture is worth a thousand words, but a picture that also contains twenty words communicates a lot more than a picture containing only reference numerals.  The stinker being that the person who is reviewing the patent would prefer that the twenty words be in the person’s own language.   A searcher in the US would prefer that the twenty words be in English.  A searcher in China would prefer that the twenty words be in Chinese.  And so on.

I have received Office Actions in USPTO cases where the Examiner has required me to insert words into a figure that had only reference numerals.  Of course the words were required to be English words.

Practitioners I know in countries outside the US have likewise received Office Actions where the Examiner has required local-language words to be inserted into a figure.

What is needed, I suggest, is a markup language (an extension of XML) for patent figures.  There are markup languages for chemical formulas and chemical structures and for math formulas;  indeed the chemical and math markup languages have been explored for use in XML-formatted patent applications.  The chemical and math markup languages were explored by USPTO back during USPTO’s failed effort to force applicants to file US patent applications in XML (the ePave system from many years ago).  The chemical and math markup languages were explored by WIPO and EPO for e-filing XML in both of those Offices.  Even now it is possible to e-file a PCT patent application in XML and this will get you an extra fee reduction of 100 Swiss Francs in addition to any other fee reductions that you might get from other e-filing accomplishments.

The idea of using a markup language is that it is all character-based, and smart computer programmers can find ways to automate downstream processes based upon the markup data.  With an appropriate markup language, the patent filer could provide an image and anchor points within the image where text is to appear.  And then the text would be provided in an XML format.  A rendering engine would vector-draw the text into the figure, and could carry out machine translation as needed prior to the rendering.  This would then permit rendered drawings in any desired publication or filing language.

I have looked around a bit and I am unable to find such a markup language already in use or in development.  But if there were to be such a markup language, it would be very friendly for an applicant that plans to file in multiple countries each of which (understandably) wishes to see its own language in the drawings.

So the comments that I envision might be along the following lines:

  • A comment supports continued efforts to migrate PCT filers toward use of e-filing software, recognizing that the software validates a proposed Abstract and warns the filer if the Abstract exceeds the word count
  • A comment recognizes that the present-day software (quoted above) prompts the user to provide in computer-readable form the text that appears in the figure that is proposed for the front page of the publication
  • A comment does not object to a relaxation of the requirements that the ISA review the Abstract for its substantive content
  • A comment observes that if a markup language could be developed for patent drawings containing text, this would offer advantages not only for Offices but also for applicants, as it would make it easier to provide patent drawings in a local language to satisfy local filing needs and local search needs

A chief issue is whether my reaction is widely held or not — is it really true that in today’s world the Abstract doesn’t matter much for searching purposes?  Maybe some would say that they do think the Abstract continues to this day to be important for searching purposes.  What do you think?  Should the ISAs be required in future to carry out detailed reviews of Abstracts as they have in the past?  Or is this not really needed any more?

By way of background, keep in mind that the PCT rules and guidelines have never required the ISAs to review, for example, the form or substance of the “background of the invention” portion of a patent application or the “summary of the invention” portion of a patent application.  Somehow the progress of science and the useful arts continued even with ISAs not reviewing those portions of patent applications.  So I invite us to consider that even if there was once a reason for ISAs to spend time reviewing Abstracts, maybe that is just not needed any more.  Do you agree?


Please post your comments below, and if you’d like to join some other practitioners in developing comments from the listserv, please make your interest known at the the PCT listserv.

One Reply to “Does the Abstract matter any more?”

  1. I also remember doing searches in CP3&4 — what I don’t remember is how I made copies of the relevant prior art that I found — copy machines at a nickel a page?? Probably ….. and I was also a Metroliner rider & didn’t realize ’til you pointed it out that the train I take today is slower ……

    I must admit that I still read the Abstracts during some searches. When I’ve performed a “first tier” search and winnowed down the world to a list of 100+ “interesting patents”, I’ll then skim through the listing (using FreePatentsOnline) that gives me the Abstract to review — I can usually toss out the 95% (hopefully) irrelevant art at this second stage. Would I be upset if Abstracts went away? Maybe — I still use the English language translations of foreign language cited art to see if I need to spend the $$ to get a full translation of a reference ….

    just some more food for thought …

Leave a Reply

Your email address will not be published. Required fields are marked *