The dead line

There is something ominous about the word “deadline” in English.

It makes you suspect some dire background. And rightly so, according to some online resources, which mention a line that you would cross at the risk of losing your life.

The French word “délai” sounds much less sinister, but the consequences of missing a “délai” are pretty much the same as those of missing a “deadline“, as a recent decision of the Paris Cour d’appel reminds us.

In this case, a French patent application was filed on November 6, 2015, claiming the benefit of the filing date of an earlier French application, namely April 1, 2015. This is what is sometimes conveniently referred to as an “internal priority claim“.

One of the requirements of claiming an internal priority is that you must provide a copy of the initial patent application (plus you must identify the modifications made in the later filing). As the applicant failed to do so upon filing the later application, the INPI issued a communication dated August 9, 2016 and notified on August 16 to the representative, inviting the applicant to submit the missing copy within a two-month period.

The applicant still did not file the requested document, so that a decision to refuse the application was issued on November 4, 2016, and notified on November 10, 2016 to the representative.

On January 8, 2018, the applicant filed a request for restoration of its rights. On August 17, 2018, the INPI issued a negative decision, holding that the request for restoration was inadmissible as it was filed past the statutory deadline.

The applicant lodged an appeal in front of the Paris Cour d’appel, which resulted in a judgment dated November 5, 2019.

The key provision in this case is article L. 612-16 Code de la propriété intellectuelle:

The applicant who failed to comply with a deadline set by the [INPI] may file a request for restoration of its rights, if it has a legitimate excuse and if the non-compliance directly resulted in a refusal of the patent application or of a request, in the lapse of the patent application or patent, or in the loss of any other right. 

The appeal must be filed […] within a two-month deadline from the removal of the cause of non-compliance. The missing act must be performed within this deadline. The request is admissible only within a one-year deadline from the expiry of the missed deadline.

[…]

The main question is that of the starting point of the one-year deadline. The INPI currently considers (and this is reflected in its Guidelines for examination, section E.3.3.1.b, Oct. 2019 edition) that the “missed deadline” starting point is the “initially missed deadline”.

As a side note, this did not appear in the previous versions of the Guidelines – I suspect that they may have been modified based on the present case.

Here, the applicant was notified on August 16, 2016 of the invitation to file the copy of the earlier filing. The deadline for filing this copy expired on October 17, 2016. The INPI thus deemed that the one-year restoration deadline expired on October 17, 2017.

The talking calendar – the must-have gadget for handling your deadlines this year.

The applicant’s defense relied on the fact that there was another possible legal remedy, namely a request for further processing. See article R. 612-52 Code de la propriété intellectuelle:

If a patent application is refused or may be refused due to non-compliance with a time limit set by the [INPI], the refusal is not issued or is revoked if the applicant files a request for further processing. The request must be filed in writing within a two-month deadline from the communication of the refusal decision. The missing act must be performed within this deadline. The request is admissible only if the required fee is paid. 

Thus, the applicant argued that the missed deadline was the two-month further processing deadline, and not the initial two-month deadline starting from the INPI communication. The request for restoration was thus filed with respect to the further processing deadline.

This reasoning will probably sound familiar to many European readers, as this is exactly the approach used at the EPO.

Applying this reasoning to the facts at hand, the deadline for requesting further processing expired on January 10, 2017 (date of notification of the refusal decision + 2 months), and therefore the one-year restoration deadline expired on January 10, 2018 – so that the request for restoration was timely filed.

But the appeal judges did not follow this argumentation. They stuck to the INPI’s interpretation of L. 612-16 and considered that the missed deadline used as a starting point must be the initial missed deadline, and not the further processing missed deadline. The INPI’s decision was accordingly confirmed.

Don’t ask me for the reasons why they chose one interpretation over the other. They just did, and there is no actual justification in the judgment.

This is all the more unfortunate as the summary of the facts reveals that two interesting case law decisions were discussed by the petitioner: a cassation ruling dated April 15, 1986, and a later judgment of the Paris Cour d’appel dated January 14, 1987.

The petitioner argued that the INPI misinterpreted the 1986 cassation ruling – from which it can be inferred that this ruling was the basis for the INPI’s position.

The facts underlying the 1986 ruling are that a patent application was refused for failure to pay the grant fee; a request for restoration was filed but it was held inadmissible as it was filed more than one year after the expiry of the time limit for paying the grant fee. The petitioner tried to take advantage of an additional two-month period corresponding to the filing of a request for further processing – to no avail.

The cassation judges stated the following: 

The non-compliance of the deadline for the request for further processing is not excluded from the provisions of [what is now art. L.612-16]. But the provisions of [what is now art. R. 612-52] do not result in extending the one-year deadline set in the second paragraph of [what is now art. L.612-16]. The Cour d’appel held that, irrespective of the grounds for the request for restoration, it is only admissible within a one-year deadline from the final date on which the initially omitted act had to be performed, and therefore rightly justified its decision. 

This is quite clear, right? The one-year deadline must be computed from the expiry of the initial deadline. Well, not so fast.

Looking at the specifics of the case that led to the 1986 cassation ruling, it turns out that, in that case, the request for restoration had been filed with respect to the initial time limit (i.e. the time limit for paying the grant fee). It had not been filed with respect to the (missing) request for further processing. 

Let’s now turn to the second case discussed by the petitioner. It is an appeal judgment which was issued after the 1986 cassation ruling. In this case, it was also the grant fee which was not paid in time. But the difference is that the applicant filed a request for restoration with respect to the (missing) request for further processing.

And the appeal judges computed the one-year deadline from the expiry of the period for requesting further processing, not from the expiry of the initial deadline for paying the grant fee: 

[…] The petition is admissible as to time limits. Indeed, according to [what is now art. L.612-16] the time limit for filing a request for restoration is one year from the expiry of the deadline not complied with. The request for further processing was due by April 17, 1982, so that the time limit for filing the request for restoration expired on April 17, 1983. The request for restoration was filed on April 15, 1983.

In summary, reading between the lines, we have:

  • On the one hand, a 1986 cassation ruling which could at first sight be seen as supporting the INPI’s position, except that the facts were different and that the relevant statement in the ruling is ambiguous. There is a reference to “the initially omitted act” but it is unclear what we should make of it. The cassation judges did simply not have to decide on a request for restoration with respect to a request for further processing.
  • On the other hand, a 1987 appeal decision based on facts similar to the present facts, and that is manifestly contradictory with the decision commented on today.

It is thus a real pity that the Cour d’appel did not seriously address these issues. 

I would also add that, one a purely pragmatic standpoint, it seems to me that there may be some contradiction in the INPI’s approach (and in the court’s present approach) to this matter. Indeed, the INPI acknowledges that a request for restoration may be filed with respect to a missed deadline for a request for further processing. This is stated in the table that you can find on the last page of the Guidelines for examination.

But what is the point in requesting restoration of right with respect to the further processing deadline, if the one-year time limit is to be computed from the initial missed deadline? I cannot think of any situation in which requesting restoration with respect to the further processing deadline would then be of any use.

I would not like to conclude this post without wishing all readers a very happy new year.

I hope that I have not missed the deadline for doing so – but hey, I have always heard that it is all right to send your best wishes until January 31. This may not be true elsewhere, but as you can see there is some cultural specificity about French deadlines.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 1, November 5, 2019, … v. Directeur Général de l’INPI, RG No. 18/20057.

Divide and not conquer

Many decisions that I have commented on relate to complex cases, involving multiple patents, or various actions in front of different courts, or endless back and forth bouncing between the Cour de cassation and the Cour d’appel.

So I am grateful, that, just this once, the mailman has delivered a short decision on a relatively simple legal point.

On March 21, 2008, Kubota Corporation filed a French patent application on a lawn mower, published as FR2915847.

On April 22, 2015, they filed a first divisional patent application, published as FR3019966.

On March 1, 2018, they filed a second divisional patent application. However, on August 27, 2018, the INPI (French patent office) refused this second divisional application as inadmissible.

The reason for the refusal was that the second divisional application was filed after the fee for grant for the grandparent application (FR’847) had been paid, back in April 2015.

Indeed, the INPI Guidelines for examination recite the following (section B.III.2):

A division can only be requested until the date of payment of the fee for grant and printing for the initial application. When this fee has already been paid or when the grant procedure has ended because of a refusal, withdrawal or lapse of the initial application, the division is refused. […]

When the division relates to a divisional application (B) stemming from an initial application (A), this control concerns application (A), and not divisional application (B)

French practice therefore differs from EPO practice.

At the EPO, you may file a second generation divisional application as long as the first generation divisional application is still pending, even if the grandparent application has already proceeded to grant. This is not possible according to the INPI guidelines.

Kubota filed an appeal against the refusal with the Paris Cour d’appel. They challenged the INPI’s interpretation of the statute.

The provision at stake is article R. 612-34 of the Code de la propriété intellectuelle:

Until the payment of the fee for grant and printing of the patent, the applicant may, on its own initiative, file divisional applications of its initial patent application. 

The key issue is how to understand to which application the time limit refers. The INPI’s position is that it is the original application in the family; whereas the appellant’s position is that it is rather the application from which a division is filed.

In this case, the fee for grant and printing for the first divisional application (FR’966) was paid on March 7, 2018, i.e. after the filing of the second divisional application at stake here – so that the filing would be admissible under the appellant’s interpretation of R. 612-34.

The court sided with the INPI on this interpretation issue.

First, they made reference to article L. 612-4:

The patent application can only relate to one invention or a plurality of inventions which are related so as to form a single general inventive concept. 

Any application not meeting the requirements of the previous paragraph must be divided within the statutory deadline; divisional applications benefit from the filing date and optionally priority date of the initial application

Thus, the court concluded that the “initial application” is the first one in the family – the one that defines the filing and priority dates for the entire family.

Second, the court stated that, in article R. 612-34, the term “patent” (cf. “until the payment of the fee for grant and printing of the patent”) refers to the same IP right as the “initial patent application” (cf. “may […] file divisional applications of its initial patent application“).

Ergo, the time limit condition applies to the payment of the fee for grant and printing in connection with the original application in the family.

In its submissions, Kubota made reference to articles L. 612-12-3 and L. 613-25 which prohibit extension of subject-matter in a divisional application, beyond the content of the “initial application as filed“. But the court was not convinced, noting that, precisely, this prohibition applies to a comparison with the original application in the family.

The court also remained insensitive to a comparison with other patent offices having a different practice: “the harmonization objective relates to the principle of the ability to divide a patent application, not the modalities of this division”.

All in all, the court’s position does not appear unreasonable. But is it compelling?

This lawn mower looks skeptical about the appeal decision.

Let’s take a step back. Article R. 612-34 does not strike me as being crystal clear. I suspect that this may be because its drafters did not have in mind the case of cascading divisional applications, i.e. divisional applications of divisional applications.

Furthermore, both main points of the court’s reasoning could potentially be challenged.

Starting with the second point in the court’s reasoning, how can we be certain that the term “patent” and the term “initial patent application” in R. 612-34 relate to one and the same thing? After all, they are not the same terms.

And then, going back to the first point, the court referred to L. 612-4 and concluded that the term “initial patent application” refers to the original application in the family.

But: 1, even if this is the case, the term “initial patent application” may or may not have been used perfectly consistently in the Code de la propriété intellectuelle; where is the evidence that it has exactly the same meaning in R. 612-34? And 2, “initial patent application” in L. 612-4 could also be interpreted as relating to the (immediate) parent application based on which the division is performed. It is true, that ultimately, the same filing date / priority date applies to the entire family, but in the case of cascading applications, it could simply be because the filing date of the “initial” application (meaning the parent application) is itself defined by reference to the filing date of its own “initial” application (meaning the grandparent application). A transitive definition, if you will.

Again, article R. 612-34 recites that “the applicant may, on its own initiative, file divisional applications of its initial patent application“. On the face of it, isn’t the “initial patent application” in this provision the one based on which the division is made, i.e. the immediate parent?

If so, we would come to the conclusion that:

  • Either it is only possible to file a divisional application based on the original application, i.e. cascading or second generation divisional applications are not allowed.
  • Or else, cascading applications are allowed, and the time limit in R. 612-34 applies to the previous application in the cascade, not to the source of the cascade.

In turn, if the above reasoning is correct, then this means that the INPI’s position that has just been validated by the court (cascading divisional applications are OK but the time limit applies to the original application) is contradictory.

In fact, on a practical standpoint, what is the point in allowing cascading divisional applications if you can only file a divisional application when the original application in the family is still pending? This is rather useless, and you may as well decide that only divisional applications stemming from the original application are possible.

As a final and important remark, the court acknowledged in passing that the INPI’s practice prior to 2011 was different, but added that this does not matter.

Well, does it not? I am quite curious about this. Do readers have any information to share about this change of practice and what caused it? Article R. 612-34 has been here and has not been modified since 1995.

Now that it is time to finish this post, I am left with the impression that even simple legal points are not necessarily straightforward.

Let’s see if there is a cassation appeal, and if so, what the court will make of this.


CASE REFERENCE: Cour d’appel de Paris, pôle 5 chambre 2, November 22, 2019, Kubota Corporation v. Directeur Genéral de l’INPI, RG No. 18/27433.

A technical silver lining

They say that every cloud has a silver lining.

Thinking about a recent Philips case in which not one, not two, but three patents were revoked by the Paris Tribunal de grande instance (TGI), this saying seems quite fitting.

From the perspective of IP right holders, the dark cloud is of course that we have just been presented with yet another confirmation that patent validity is severely appraised in France. But the silver lining is that, although software patents were at stake, the court did not strike them down for just being that, but on the contrary fully acknowledged the technical nature of the inventions.

The case at hand was certainly a big one for Philips.

Not only were three patents involved, but there were also parallel proceedings against several unrelated defendants.

So much so that there were in fact four judgments issued on the same day: one pitching Philips against HTC, another against Asus, another against Archos and the last one against Wiko.

Philips reproached all defendants for not joining their “Touch enabled device” license program, and retaliated by filing an infringement action based on European patents No. EP 1384134, EP 0888687 and EP 1571988.

The four judgments look substantially identical, so it should be sufficient to look at only one of them.

Claim 1 of EP’134 reads as follows:

A data processing apparatus having a user interface assisting in searching for information from an ordered list a data array, the apparatus comprising:
an array scroller responsive to user actuation; and
a helper character-generator, actuated by continued user actuation of the array, the helper character generator being operative to display a helper character representative of a portion in the list being scrolled.

In brief, this is a user interface invention. The context is that of a long list of elements. In order to facilitate the retrieval of one element in the list by the user, a so-called “helper character” appears. This helper character may represent e.g. a first letter of a name or a first digit of a telephone number. It is thus representative of a cluster of entries in the list. By scrolling from helper character to helper character, the user may select a desired cluster, and then browse within this cluster instead of browsing through the entire list.

Choosing the right song in a menu, a critical decision to make.

The defendants argued that this was a mere presentation of information, and thus not an invention under art. 52 EPC.

Here is the court’s position on this objection:

Displaying the helper character involves continuous actuation by the user in a scrolling array. 

The patent thus relies on the user taking an active part, via an improved interface interacting with him, thus creating a particular technical effect due to the actuation by the user, allowing a faster choice in an array of elements. 

By combining several modes of scrolling down a list of elements, the patent teaches an adaptive user interface for searching and displaying, and thus does not just present information “as such”. It should also be emphasized that the purpose of the invention is to facilitate the selection of an element in a large set of data, on small size devices, taking into account their specific ergonomic design. 

Therefore, the EP’134 patent undeniably implements technical means to obtain a technical solution to a technical problem. Its subject-matter is thus patentable. 

I asked my partner Aujain Eghbali, who is an ace at computer-implemented inventions, what we can make of this.

His main comment was that this is the same type of reasoning that the EPO could have made. He added that the court’s reasoning is in fact quite patent-friendly even by EPO standards.

In particular, the court relied on the ergonomic character of the human/machine interface when assessing the technical nature of the invention. The Boards of appeal of the EPO have sometimes adopted a similar approach (see T 1779/14) but the Guidelines for examination do not take any clear position on the technical nature of such innovations. Section G-II, 3.7.1 suggests that a “physical ergonomic advantage” could be considered as a technical contribution, but specifically in relation with user input.

The court also took note of a statement in the description of the patent, per which human-machine interaction is particularly challenging when the device is small, due to the physiology of the user. Although effects relying on human physiology are acknowledged in the EPO Guidelines (section G-II, 3.7), they are not commonly accepted by examiners as having a technical character, Aujain said.

Going back to the judgment, the EP’134 patent was thus found to meet the requirements of article 52 EPC.

Further objections of insufficiency of disclosure and lack of novelty were also rejected. But the patent finally died of lack of inventive step.

However, it is important to note that all features of claim 1 were fully taken into account in the inventive step reasoning, in keeping with the court’s finding that the claim related to a technical solution to a technical problem.

The court’s negative opinion on the patent was thus fully based on relevant disclosures identified in the prior art. So, this is your good old, traditional, prior art-based lack of inventive step, if you will; not a this-is-not-technical-enough lack of inventive step.

In more detail:

The EP’036 document teaches that, upon continuous actuation of the scrolling key, the first names are sequentially displayed, in the alphabetic order, for each letter, thus providing a search tool. 

Starting from this search tool disclosed in EP’036, the skilled person, namely an electrical engineer experienced in designing user interfaces for portable electronic devices, confronted with the problem of selecting an element in a scrolling list on a small size screen, would achieve claim 1 of the EP’134 patent, since the search tool of the EP’036 prior art (underlining the first letter of a name) is only a visual alternative to the generation of a help character, which is anyway taught by the US’949 document. 

The EP’036 document also taught scrolling cluster by cluster, allowing the user, by maintaining the scrolling key pressed, to go from letter to letter in the directory, without having to go through all names starting with the same letter. 

Therefore, the features of claim 1 of the EP’134 patent derive from the state of the art. 

Next up was EP’687, another user-interface patent.

See in particular claim 1:

An electronic device comprising:
– at least one display;
– a controller arranged to cause the display to show a rotating menu comprising a plurality of menu options,
which menu is disposed off centre in the display so that at least one option is rotatable off the display at any one time, whereby an arbitrary number of options may be added to the menu without changing its format.

Again, the court fully acknowledged the technical character of the invention:

[…] The EP’687 patent aims at improving a user interface for a screen. To this end, it allows displaying a menu comprising a plurality of options as well as adding an arbitrary number of options to the menu without changing its format. It thus provides a better identification of options. 

The objective of the EP’687 patent is to facilitate adding more articles to the menu without making the display more complex, as long as at least one option in the menu is out of the screen due to the rotating menu being off-centered. 

Therefore, the patent makes it possible for the user to interact with the device, while adding any number of options to the menu, without modifying the format and while keeping their own character making their identification easier.

Consequently, the EP’687 patent, which ensures a better identification of menu options for the user, by means of a controller which combines a particular option display format and a double function of rotation and addition of options, obviously implements technical means to provide a technical solution to a technical problem. Its subject-matter is thus patentable. 

Like for the previous patent, Philips was also able to fend off insufficiency and lack of novelty challenges, but then stumbled on inventive step, as the prior art was deemed to be very similar to the invention.

The third patent was directed to an activity monitor. Its technical character was not challenged. It was revoked for lack of inventive step all the same.

In summary, the present case seems to confirm a trend recently seen in a Thales decision and on the contrary to drift further away from past case law in which the stance on patent eligibility was considerably stricter.

How apropos then that the INPI (French patent office) has just published an updated version of its examination guidelines containing two new sections C.VII.1.3.1 and C.VII.1.3.2 on computer-implemented inventions.

One section is dedicated to CAD and computer simulations. The other one relates to AI. Both provide additional guidance as to what should be considered technical or non-technical, again quite in line with the EPO approach. Maybe the next revision will contain some additional thoughts on graphical user interfaces?


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 1ère section, July 11, 2019, Koninklijke Philips NV v. SAS HTC France Corporation & HTC Corporation, RG No. 16/02073.

Would a “could” suffice?

Comparing EPO and French case law on validity is one of this blog’s favorite topics.

Today’s post will be dedicated to a cornerstone of the EPO doctrine: the could/would approach in the inventive step reasoning.

The could/would approach was actually nicely summarized by the Paris Cour d’appel in a Sandoz v. Eli Lilly decision (January 13, 2012), making explicit reference to the European case law book that all European patent attorneys have on one of their shelves:

As can be derived from established case law of the Boards of appeal of the EPO, the question is not whether the skilled person could have made the invention by modifying the state of the art, but whether he would have done so in the hope of achieving the advantages that were really obtained in the light of the technical problem at stake, because the state of the art contained suggestions in this respect. 

So far so good, and all supporters of pan-European harmonization must be satisfied.

Not so fast, though, as some recent decisions seem to show that the could/would approach is not as inscribed in the DNA of French IP courts as it is in the EPO’s DNA. Here is an example. 

Modelabs Mobiles is a French company which distributes mobile phones and other IT devices. Modelabs Mobiles filed a French patent application No. FR 3063299 on February 24, 2017, directed to a casing for locking and immobilizing a phone.

Just after the publication of the application, Modelabs Mobiles proceeded with an infringement seizure against its competitor Wiko, on October 23, 2018, and soon thereafter filed an infringement suit, based on the FR’299 application (yes, this is possible before the grant of the patent under French law).

The patent was granted on April 12, 2019. On June 7, 2019, Modelabs filed a claim for preliminary injunction with the judge in charge of case management, and further requested a provisional damages award of more than 2 million euros. Modelabs argued that the existence of infringement acts was likely, and that since the defendant was in a restructuring process, there was a risk that damages could not be recovered.

The judge in charge of case management ruled on September 13, looking first at whether the patent was apparently valid.

Here is claim 1 of the patent:

Casing, in particular intended to contain a telephone, comprising:

– a first shell and a second shell that close onto one another so as to close the casing by rendering its contents inaccessible,

– a slide that is movable in translation on the first shell, comprising at least one coupling member, and which can be actuated from outside the closed casing so as to be positioned by translation in a first, a second or a third positions, wherein the first shell, the second shell and the coupling member(s) are arranged so that, when the casing is closed:

– at least one coupling member is coupled to the second shell when the slide is in the first position, thus opposing the opening of the casing,

– at least one coupling member is coupled to the second shell when the slide is in the second position, thus opposing the opening of the casing,

– the coupling member(s) is/are free when the slide is in the third position, thus allowing the opening of the casing, the casing further comprising:

– a device for locking the slide on one of the shells in the first position, actuable only by means of a dedicated tool from the outside of the casing when this casing is closed,

– first means for reversible immobilization of the slider in the second position,

– second means for reversible immobilization of the slider in the third position.

In summary, this casing can be used for protecting a phone in a store. The casing has three configurations: one which is fully locked, and which requires a tool (e.g. a magnetic one) to be unlocked; one which is unlocked and closed; and one which is unlocked and open. The transitioning between the unlocked / closed and unlocked / open positions can be performed without using an external tool.

Locked – unlocked

Wiko relied on three prior art documents against the novelty and inventive step of the main claim. The judge found that the claim was novel.

Regarding in particular the so-called EP’090 prior art reference, the judge held:

[…] The “movable member” which can correspond to the “slide” of the invention of the patent can be positioned in two different positions, and not three. In order to ensure an anti-theft function, a safety device must be added, comprising an “inserted” external element whereas, in the casing claimed by Modelabs, the safety device is fully part of said casing even if it requires the use of an external tool to be deactivated. 

A similar conclusion was reached regarding the two other prior art references.

However, regarding inventive step, the judge made the following comments:

In view of the state of the art […], the skilled person already knew casings for products on sale, having an integrated locking system, which can be actuated by a dedicated external tool such as a magnetic tool, preventing access to the content of the casing in the absence of said tool without breaking the packaging itself. 

Admittedly, the casing at stake, relative to [this prior art], provides an intermediate position in addition to the mere locked/unlocked alternative. but this mode of using the casing (simply closed in a reversible manner) does not address the technical problem that the invention purported to solve, which was to make the content of the casing safer. 

Besides, it can be legitimately questioned whether the skilled person requires a particular inventive step in adding an intermediate position between the locked/unlocked positions for the coupling member supported by the slide […], in order to keep the casing closed in a reversible manner (thus that can be manually actuated, without a dedicated tool), in view of the state of the art. Indeed, the addition of such an intermediate position prima facie seems to rather logically derive from [the prior art]. 

Therefore, the prior art submitted as a defense provides a serious possible ground for nullity [of the patent]. 

As you can see, the judge does not seem to reason in terms of what the skilled person would have done, but rather in terms of what he or she could have done.

Indeed, although there does not seem to be any suggestion or motivation in the prior art to have this intermediate closed/unlocked position for the slide, as far as I can tell from the judgment, the judge nevertheless concluded that it would have been “logical” for the skilled person to include such an additional position.

By way of comparison, here is the position taken by the EPO examiner in charge of the search report for the French application (based on different, but presumably similar prior art):

Such a casing with three positions for the slide and “reversible immobilization means” […] is completely unknown in the state of the art. Hints pointing to this direction are also absent. 

So, what can we make of this? One interpretation of course is that the inventive step threshold is higher in France than at the EPO. This is of course very possible. Another interpretation is that the judge had questions or doubts and simply did not want to grant a preliminary injunction and preferred to let the case continue on the merits.

That said, my attention was also drawn by the following statement:

Modelabs did not provide any argumentation regarding the inventive step of its patent which differs from the novelty argumentation. 

As always, we should keep in mind that, in view of our rules of civil procedure, and based on a longstanding tradition, French judges limit the scope of their decisions to the arguments put forward by the parties. Therefore, in the absence of any reference in the patentee’s submissions to the could/would approach, it is unsurprising that the judge did not use this reasoning.

Thus, I could say (or make that rather: I would say), that argumentation in French patent litigation has to be complete, including explanations on the legal standards to be applied – or else… 

As a next step, if the case continues on the merits, we will see whether the three-judge panel of the court uses the same or a different approach.


CASE REFERENCE: Tribunal de grande instance de Paris, 3ème chambre 3ème section, ordonnance du juge de la mise en état, September 13, 2019, SAS Modelabs Mobiles v. SAS Wiko, RG No. 18/13517.

No way back

There once were submarine patents, living a quiet and inconspicuous life underwater, only to surface one day with a resounding splash.

Turns out what we do sometimes have in France is submarine patent cases. A recent ruling by the Lyon Cour d’appel could be considered as one of those.

In fact, Paris has exclusive jurisdiction over patent matters since November 1, 2009. Thus, I was more or less convinced I would never see again a patent revoked in a provincial court – but I was wrong.

Chavanoz Industrie, proprietor of European patent No. EP 0900294, filed an infringement suit against Mermet SAS in 2009, a couple of months before the November 1 date, and thus in Lyon. The first instance judgment was issued in September 2016.

I assume that the main reason why this lingered for such a long time is that the defendant merged in 2011, and then underwent insolvency proceedings in 2012. In 2013, a so-called continuation plan was ordered.

At any rate, the Lyon TGI hit late, but it hit hard, as the patent was held valid and infringed, and a damages award amounting to a total of more than 25 million euros was issued.

However, the ruling was fully overturned on appeal, a few weeks ago, as the Cour d’appel had a different appreciation of a key aspect of the case from the TGI.

The patent at stake was filed on April 16, 1997, claiming a French priority of May 7, 1996. It was granted in 1999 and limited by the French patent office (INPI)… in 2017, just a few days before it expired. Unusual enough, but wait until you read about the rest of the case.

Claim 1 of the patent as limited reads as follows:

Composite yarn for making a technical fabric for an awning, said yarn comprising a core composed of a continuous yarn made of glass, and a coated sheath composed of a matrix consisting of at least one chlorinated polymer material, for example a polyvinyl chloride or PVC, and a fire-retarding filler incorporated into and distributed within the said matrix, characterized in that, in combination, on the one hand the fire-retarding filler comprises a ternary composition which combines an oxygenated antimony compound, for example antimony trioxide, a hydrated metal oxide, the metal of which is chosen from the group consisting of aluminium, magnesium, tin, zinc and lead, for example an alumina hydrate, and a zinc borate and, on the other hand, together with the said ternary composition, the total weight content of inorganic matter in the yarn is between 4% and 65%.

The alleged infringer asserted that the invention had been made available to the public before the priority date of May 7, 1996, because the patent proprietor Chavanoz had supplied the patented yarn to (among others) a third party, Helioscreen.

Chavanoz responded by relying on a confidentiality agreement signed with Helioscreen on April 9, 1996.

The court had to deal with three thorny issues:

  • Chavanoz stated that most of Mermet’s exhibits relating to the alleged public prior use were illegally obtained, in view of the confidentiality agreement, and that they should thus be disregarded by the court.
  • Then there was the question of whether the supply of the yarn in question was indeed a public disclosure or whether it was confidential.
  • Then there was the question of whether the skilled person would have been able to arrive at the claimed invention by analyzing the supplied yarn, at the priority date.

First things first, the fate of the exhibits.

In support of the public prior use allegation, the infringement defendant filed a number of documents, such as invoices, orders, letters, between the patent proprietor and the third party Helioscreen. These documents were, as far as I understand, willingly provided by Helioscreen. Chavanoz protested that this was in breach of the Chavanoz-Helioscreen confidentiality agreement of April 9, 1996.

The first instance judges agreed and thus discarded the evidence. But the appeal judges came to a different conclusion.

The 1996 confidentiality agreement recites that the parties have been cooperating since May 9, 1994 in the business of technical fabric. Helioscreen provided specifications, Chavanoz made yarn and supplied it to Helioscreen, together with some technical information, Helioscreen made fabric with this yarn and supplied samples of said fabric to Chavanoz for testing. The agreement then basically states that all information, documentation, products exchanged between the parties is confidential, and that the agreement itself is confidential. The agreement also mentions that it is retroactive back to May 9, 1994 and that it will remain in force as long as all of the confidential information has not been made available to the public.

The appeal judges held that the confidentiality agreement expired in November 1997, when the patent application was published. Therefore, Helioscreen was free to provide any documents and information to Mermet. Besides, Mermet was anyway not bound by the confidentiality agreement and was free to provide any material obtained from Helioscreen in order to build its defense. In fact, Helioscreen was not sued by Chavanoz for breach of the confidentiality agreement.

As a result, all of the exhibits filed by Mermet were deemed admissible.

I am not necessarily convinced by the argument that the confidentiality agreement expired when the patent application was published. Indeed, the information exchanged by the parties could possibly extend beyond the contents of the patent application. Notwithstanding, it seems only fair that Mermet, as a third party to the agreement, should be allowed to rely on whatever evidence is at its disposal as a defense in the infringement suit.

This brings us to the second point, namely whether the exchanges between Chavanoz and Helioscreen were confidential or not.

The court analyzed in detail the evidence at hand, and concluded that several dozens of tons of so-called M1B1 yarn were sold by Chavanoz to Helioscreen before the priority date of May 7, 1996. The yarn was first experimented in 1992, but by 1995, the production of the yarn was on an industrial scale. The court further noted that there is no evidence that the formulation of the yarn changed as from 1995-1996, or that the product was still in development – even though further yarn colors were later designed.

Besides, the M1B1 yarn was also supplied to other fabric manufacturers, including Mermet itself and another company called Brochier, in the context of the professional association ScreenGlass, in order for all manufacturers to provide feedback on the yarn.

The court also established that the fabric made by Helioscreen from the M1B1 yarn was sold to various clients before May 7, 1996.

No written confidentiality agreement was in place when all these exchanges took place; in particular, no exhibit on file bears the mention “confidential“.

The court then stated that “there is never a presumption of confidentiality in the context of a relationship with a buyer“.

Next, the appeal judges turned to the confidentiality agreement of April 9, 1996, which, as a reminder, states that all prior exchanges with Helioscreen were confidential. The court was unimpressed by this statement:

The reality of the amounts supplied to Helioscreen and the reality of the industrialization phase emphasized by Chavanoz Industrie itself in its commercial correspondence are never mentioned [in the agreement]. The content of the deliveries that are mentioned only relates to “samples”. 

[…]

It is thus important to determine if a company that has already disclosed its invention to a partner not subjected to an obligation of confidentiality at the time of the disclosure can erase the existence of this disclosure, which may be novelty-destroying, by signing a confidentiality agreement after the facts. 

The public policy provisions of patent law prevent parties to an agreement, from depriving a disclosure which has already occurred of legal effect, even with a retroactive effect.  

An invention is made available to the public when it is disclosed to a person not bound by confidentiality at the time of the disclosure, so that the retroactive provision in the above confidentiality agreement has here no effect on the appraisal of the validity of the patent. 

Is it possible to go back to the pre-disclosure situation?

Long story short, the particulars of the supply of M1B1 yarn by Chavanoz were deemed incompatible with the notion of an implicitly confidential sampling. And a public disclosure cannot be turned confidential ex post.

There was a side argument by the patentee that Helioscreen’s sale of fabric to their client Krüland was a non-prejudicial disclosure under article 55 EPC. This was considered as irrelevant by the court, since overall the public prior use took place even before 6 months prior to the priority date.

The third point at last. Chavanoz argued that the skilled person would not have been able to know the composition of the M1B1 product at the priority date without undue burden.

The defendant performed a number of experimental tests to counter this theory. First, they made the patented product according to the method taught in the patent, and then had the product analyzed. Second, they had a bailiff collect samples of the M1B1 yarn dating back to 1994 in Helioscreen’s premises, and had those analyzed as well.

The manufacturing was performed in the presence of a bailiff. All analyses were conducted in two independent labs, using analysis methods available at the priority date. Various experts’s reports were produced on both sides, either to challenge the results of the analyses, or to support them.

I will not go over the details of the discussion, which was technically quite complex, but it will suffice to say that the court was ultimately not persuaded by Chavanoz’ criticisms, and accepted Mermet’s conclusion that the M1B1 yarn could indeed be determined by the skilled person as falling within the patent claims.

As a result, claims 1, 2, 3, 5, 7 and 8 of the patent were deemed to lack novelty and were revoked, and Chavanoz’ infringement claim was dismissed.

Let me say that this decision is overall well reasoned and makes for an excellent read. I am therefore looking forward to the next submarine from Lyon – if any.

As a take home message, confidentiality agreements should really be put into place initially. There is no way back once a business relationship has started on a non-confidential basis.


CASE REFERENCE: Cour d’appel de Lyon, 1ère chambre civile A, September 12, 2019, SAS Mermet et al. v. SARL Chavanoz Industrie, RG No. 16/06896.