The Preliminary Part and especially Part 1 sought primarily to identify and analyse the differences and similarities between the two bodies of rules. Now it is important to determine the consequences of this situation for Quebec. We can say at once that, despite the similarities mentioned, the differences related to writings are major and present a problem that is all the more “exasperating” since it is the only one that is truly difficult to solve and since the Convention adopts a “take it or leave it” approach by not allowing reservations to be made (article 22). We will therefore deal with these two aspects in Section 1 of this second part. In Section 2, we will then discuss the recommendation itself.1 – Irreconcilability of Quebec Law and the Convention
 Reconciliation therefore appears to be difficult for the concept of “writing” and only for this concept, since, without having to say much more about it, we believe that the distinctions that may arise with regard to the concept of “original” can be tolerated. First of all, the basic criterion in both cases is integrity. Next, it may be thought that, in the very rare cases that involve functions 2 and 3 described in section 12 of the LFIT Act – for which we have been unable to find any examples in international documents outside the fields excluded by article 2 of the Convention – the specificity of the LFIT Act will remain applicable in light of the general nature of the Convention.1.1 – Irreconcilable “Writing”
 We have already seen that a writing materializes differently in the two sets of legal rules. The two materializations find different expressions even though the same principles are applied, namely technological neutrality and functional equivalence. This distinction can be explained in two ways. First, these instruments do not have the same function. While the Convention seeks to harmonize various national legal rules, the LFIT Act understands the concept more from the standpoint of the civil law. While the first instrument is focused outward, the second is more introspective and considers the question in light of all the writings provided for in the C.C.Q. Second, the Convention seeks to manage and, incidentally, to facilitate the transition from paper to electronic media; although the LFIT Act also has this objective, it is focused more on finding a “universal” criterion that will be applicable no matter what medium is used, whether paper or electronic. These two approaches undoubtedly explain why the respective drafters were not in the same frame of mind.
 Therefore, faced with the difficulty of reconciling the two definitions, we cannot avoid choosing between them, contrary to what we thought in Part 1. The lines that follow seek to show that, in our opinion, the integrity criterion is perfectly justified in the context of Quebec law and that adopting the “subsequent reference” criterion would impair the consistency of that law. This is justified both by the criticisms that can be made of the criterion in the Convention and by the virtues associated with the criterion in the LFIT Act.A – Criticism of the “subsequent reference” criterion
 Apart from the fact that the Convention is sometimes said to have been inspired by common law principles, several characteristics of the “subsequent reference” criterion are themselves capable of causing problems. First, the Convention interprets the concept of “writing” very narrowly and takes care to distinguish the formality associated with a “simple writing” or a writing stricto sensu from that required for a “signed writing” or “original writing”.
The requirement of written form is often combined with other concepts distinct from writing, such as signature and original. Thus, the requirement of a “writing” should be considered as the lowest layer in a hierarchy of form requirements, which provides distinct levels of reliability, traceability and integrity with respect to paper documents. The requirement that data be presented in written form (which can be described as a “threshold requirement”) should thus not be confused with more stringent requirements such as “signed writing”, “signed original” or “authenticated legal act”.
 However, such an interpretation strips all meaning from a writing stricto sensu, which, on its own, is liable to be rarely associated with a writing found in an international instrument. The Convention therefore takes great care to limit its scope. Moreover, when the existence of a writing must be determined, it will have to be ascertained whether the writing provided for in an international convention, for example, refers to a “simple writing”, “signed writing” or “original writing”. While the Convention has the virtue of providing criteria to rationalize and objectify the transition from paper to electronic media, we fear that, on the contrary, the exercise will be as ambiguous as determining “how many angels can dance on the head of a pin”.
 Second, while the concept of “writing” is narrow, the criterion for applying it, namely “subsequent reference”, is quite broad, and deliberately so, as we have already seen. It is meant to be inclusive to ensure that the largest possible number of states can become parties to the broadest possible Convention and that no particular technology is precluded from being characterized as a writing. However, in our opinion, making it too easy to create a writing is not necessarily a good thing, since this formality will have no reason to exist if it is completely distorted. This criterion might be met through a hyperlink, for example, without the link being underlined when the writing is formed.
 This brings us to a third problem with the “subsequent reference” criterion, namely that, while it undoubtedly fulfils the evidentiary function of a document, in no way does it satisfy the requirement of a formality ad validitatem, which was referred to as “being aware” during UNCITRAL’s preparatory work. A writing is often required to slow down the process and highlight the importance of a clause or an action. Even though the other contracting party can consult the contract after it is entered into, the writing requirement will not have fulfilled its formal protective function with such a criterion.
 Finally, we believe that the “subsequent reference” criterion is relatively new. To the best of our knowledge, it was thought up during the discussions under UNCITRAL. Therefore, this criterion has not existed for a meaningful length of time, and it is difficult for me to say whether it can stand the test of time. Being rather suspicious of new legal constructs is thus not pointless. It is necessary to “legislate tremulously”, a statement generally attributed to Professor Carbonnier. With all due respect, international documents, and particularly the six named in article 20 of the Convention, are quite unstable in the formal requirements they impose for writings, which include a signed writing, documents that include a telegram or telex, a “complete record . . . [that] provides authentication of its source”, the absence of any form and, finally, “subsequent reference”. This instability is especially apparent in the new version of the UNCITRAL Model Law on International Commercial Arbitration, which does not manage to identify a single solution but rather two optional ones, the first of which even raises the possibility of an “oral writing”, which will no doubt cause some interpretation problems for judges.B – Appropriateness of the Integrity Criterion in the Quebec Context
 On the other hand, the criterion of integrity, as it is dubbed in the LFIT Act, has several advantages. First, we believe that this criterion is not specific to information technologies but will apply to paper in the same way. In our opinion, as already noted, the LFIT Act does not take the same “tinkering” approach taken by the Convention. Specifically, in addition to the Convention’s desired function of reconciling paper and electronic media, the Quebec statute seeks a more “universal” criterion that can apply no matter what the medium used. Next, this criterion as described in article 2838 C.C.Q. applies only for the evidentiary function and not for the ad validitatem function. As well, it should be noted that the integrity criterion is most likely consistent with academic writing and court decisions concerning paper documents, except that the term “authenticity” was used more frequently than “integrity”, which we do not consider problematic.Finally, the criticism of inalterability in the comments on the Convention does not apply to integrity. Paragraph 145 states the following:
the concept of writing does not necessarily denote inalterability since a “writing” in pencil might still be considered a “writing” under certain existing legal definitions.
So be it. But inalterability is not integrity, since a paper document written in pencil can certainly have integrity, just like an e‑mail, whose inalterability is the same as in the previous example.
 On the other hand, neither the LFIT Act nor the C.C.Q. expressly defines a writing. And although article 2838 C.C.Q. specifically establishes the integrity criterion for four of the five categories of writings provided for in the C.C.Q., the “other writings” category does not seem clearly subject to that criterion, except perhaps through section 5, para. 3 of the LFIT Act, which also seems to choose the integrity criterion for all documents. In any event, if some doubt remains about the criterion applicable to “other writings”, that criterion can never be “subsequent reference”, which is foreign in every way to the LFIT Act and the C.C.Q.1.2 – Inflexibility of the Convention
 To begin with, it is important to note that the procedures for becoming a party to the Convention discussed in this paragraph are not part of our field of expertise. Having said this, it seems quite clear to us that reconciling the two concepts of “writing” is difficult, since article 22 of the Convention adopts a firm tone by formally prohibiting reservations:
No reservations may be made under this Convention.
 This firmness is all the more surprising given that it seems quite rare. The term is firmer than “declaration” insofar as it creates a very structured scheme set out, inter alia, in articles 19, 20, 21 and 23 of the Vienna Convention on the Law of Treaties. On the other hand, the Convention prefers the “declaration” system, a tool that seems less structured even though its effects are comparable in several respects. A state that uses such declarations cannot challenge another state’s declaration in any way.
 It seems possible to make a declaration in seven circumstances. While those related to articles 17(4) (regional economic integration organizations), 19(1) (the Convention’s date of application) and 20(2), (3) and (4) (the specific case of certain conventions) are clearly not applicable to our “writing” problem, two points can be made here. First, with regard to domestic territorial units, article 18(1) provides as follows:
If a Contracting State has two or more territorial units in which different systems of law are applicable in relation to the matters dealt with in this Convention, it may, at the time of signature, ratification, acceptance, approval or accession, declare that this Convention is to extend to all its territorial units or only to one or more of them, and may amend its declaration by submitting another declaration at any time.
 To the best of our knowledge and understanding, it will be important for Canada to assess the possibility of such a declaration for the province of Quebec. We do not think we have to make a decision on this, since it is outside our terms of reference. In either case (signing the Convention with a declaration excluding Quebec or not signing the Convention), Quebec law will not be subject to the Convention. As well, this subject remains outside our area of authority; nonetheless, it seems that this possibility applies mainly where there is a subsequent change in the law of a province, which, as we have seen, does not seem possible here.
 Second, still seeking flexibility as authorized by the Convention, article 19(2) provides as follows:
2. Any Contracting State may exclude from the scope of application of this Convention the matters it specifies in a declaration made in accordance with article 21.
 Unfortunately, the way this article is understood and the way it is interpreted in the explanatory notes do not lead us to believe that it can be used in our case.2 – Recommendation not Accede to the Convention
 The application of the Convention by Canada is a somewhat more complex question in relation to Quebec law as it now stands than it is in relation to the law of most of the common law provinces. Those provinces have generally adopted substantive and formal approaches that are modeled closely on the Model Law of 1996, on which the Convention also draws extensively. Nonetheless, in spite of differences that are not necessarily major, our analysis suggests that it would be difficult for Canada to accede to the Convention without impairing the consistency of Quebec law. In our opinion, this statement is justified both by the reasons already mentioned and also by several observations which we will now make.
 First, it is difficult for Quebec to change the integrity criterion without upsetting its own consistency. It is hard to accept that Quebec could use a writing criterion for its international transactions that differs both conceptually and in its application from the criterion used for its internal acts.
 Next, as of February 2008, 18 states have signed the Convention, the closing date indicated in the Convention being January 16, 2008 (article 16(1)). With all due respect, aside from Russia and China, none of these states seems to be a determinative partner for Canada. While the situation in the United States is no doubt more favourable to signing the Convention, the European countries seem more reluctant. Therefore, for the moment, it would seem that UNCITRAL has not been fully capable of achieving the desired consensus.
 As well, we have seen a tendency which is fairly recent, by and large in some international documents to try to simplify or even eliminate manifestations of form. In our opinion, this is reflected in the extreme formal simplicity required by the Convention for writings. However, absence of form cannot mean absence of proof, and I am not sure that formal simplicity must be associated with progress, especially in an electronic context in which more form may in fact be needed to counterbalance the absence of physical materials.
 Finally, there is a very active debate about this in the field of arbitration, among others, where the opposition between advocates of “strong” writing and those of “weak” or no writing has prevented any formal changes to the New York Convention of 1958. Changes have been achieved only through informal standards (soft law), which are quite unclear and which involve optional clauses, a fact which I believe is unique. The Convention gives the impression that it will try to find a general solution to a problem that could not be solved in the specific context of arbitration. The concept of “writing” may therefore be more complex and variable than the very simple or even simplistic definition the Convention seeks to impose.