Tuesday, March 12, 2013

Rockefeller University Press: CC-BY is not essential for Open Access

The new Open Access (OA) policy that Research Councils UK (RCUK) plans to introduce on April 1st has proved highly controversial within the research community.
Mike Rossner
 Some have expressed concern over its preference for Gold OA (OA publishing), and its concomitant disdain for Green OA (self-archiving). Others have been angered by its vacillating attitude towards the appropriate length for self-archiving embargoes.

But what may turn out to be the most divisive aspect of the new policy are its licensing requirements, notably its insistence that when RCUK-funded researchers embrace Gold OA, and pay an article-processing charge (APC), the publisher must make the paper available under a CC-BY licence.

Adding to the discomfort of those who are unhappy with this funding condition, on the same day (April 1st) the UK’s Wellcome Trust  — which was highly influential in the development of the RCUK policy — will introduce a similar rule.

The divisiveness of this new licensing approach is, perhaps, no better demonstrated than the decision by the executive director of Rockefeller University Press (RUP) Mike Rossner to pen an editorial called “New mandates? No problem for The Rockefeller University Press”.

In the editorial, Rossner questions the need for RCUK and Wellcome to insist on CC-BY, and challenges their justification that it is necessary to do so in order to permit reuse of the research they have funded, particularly through text and data mining of papers.

OA-friendly


It is important to note that over the years RUP has gained for itself an enviable reputation as one of the most (if not the most) OA-friendly of all the traditional publishers.

RUP insists on no more than a six-month embargo before the papers it publishes can be made freely available — and in fact it releases them all itself at that point. Moreover, unlike other subscription publishers, RUP licences all its content under the more liberal CC-BY-NC-SA licence.

In addition, Rossner has on a number of occasions publicly criticised other publishers when they have sought to derail OA — e.g. here and here.

Indeed, so OA-friendly did Rossner appear to be to OA advocates that in 2009 the OA advocacy group SPARC honoured him as a SPARC Innovator. As SPARC put it at the time, “Mike Rossner is an anomaly, of sorts, in the scientific publishing world. He is a force from within the establishment pushing for policies to make information more widely accessible and verifiable.”

Since then, Rossner has continued to prove OA-friendly. Last year he publicly supported the Federal Research Public Access Act (FRPAA) — which, amongst other things, would have reduced the embargo period specified in the Public Access Policy of the US National Institutes of Health (NIH) from 12 months to 6 months, and required all the major agencies of the federal government to introduce the new strengthened policy too.

Likewise, Rossner publicly disavowed the publisher-sponsored Research Works Act (RWA), a US bill that would have had the opposite effect to FRPAA, reversing the NIH Public Access policy and preventing other federal agencies from imposing similar requirements on researchers. 

And last year Rossner — with SPARC’s Heather Joseph, plus Michael Carroll, and John Willbanks — co-founded Access2Research, and launched the petition that led to the recent executive directive ordering all US Federal Agencies with research and development budgets over $100M to develop public access policies within twelve months (referred to by Rossner below as the White House directive).
 

Why?


Why then is Rossner now questioning the new OA policies that RCUK and Wellcome plan to introduce? The nub of his objection is that the CC-BY licensing requirement will serve to alienate subscription publishers, despite the fact that the same objective could be achieved in a way that is compatible with a subscription-based business model.

First, he argues, the CC-BY licence is not essential for text and data mining because copyright law’s fair use exemptions already permit such activity. And to underline RUP’s willingness for third parties to mine its own journal content (for specific purposes), the publisher recently clarified its licensing policy.

Second, Rossner says, other less controversial licences would be just as effective as CC-BY for allowing text and data mining.

Moreover, he adds, if the funders insist on the use of CC-BY, they will fragment the licensing of their own content, and confuse third parties as to what can and cannot be done with a journal’s content — since papers published as Gold OA will be licensed under CC-BY, while papers published in subscription journals, and made available after an embargo as Green OA, will be subject to a variety of different licences.

What is certain, he adds, is that no subscription publisher is going to embrace CC-BY — since it would permit the wholesale resale of that publisher’s content, and so undermine its subscription sales.

In other words, it will not be possible to standardise on CC-BY. Therefore a far better solution, says Rossner, would be for funders to mandate CC-BY-NC for all their content.

In addition, if the two funders’ actions end up having the effect of forcing subscription publishers to embrace Gold OA prematurely, it could have serious financial implications for those, like RUP, who publish “selective” journals, since for them the cost of publishing a paper is higher than any article-processing charge that these funders are likely to agree to subsidise.

RUP could, of course, offer Hybrid OA and subside its OA papers with subscription revenues, but Rossner does not want to do this. As he puts it, “I believe it sends the wrong message to the research community for a subscription-based journal to make an article free immediately for less than the actual cost of doing so. In such a case the immediate access fee is being subsidized by subscribers, who may not want to subsidize immediate access to a subset of articles.”

He sums his position up in this way: “I do support licensing terms that provide reuse rights. I just don’t think it is necessary to invoke a licence that is potentially detrimental to subscription-based publishers to do so.”

In short, CC-BY is not essential for OA, or even for text and data mining, and will alienate some of OA’s friends. So why insist on it?

Whatever one’s views on the rights and wrongs of CC-BY for research papers, it would surely be regrettable if RCUK and Wellcome ended up alienating one of OA’s greatest friends — one, “from within the establishment pushing for policies to make information more widely accessible and verifiable”.

Rossner’s editorial can be read here.

A Q&A interview conducted with the publisher by email follows below.

The interview begins …


RP: As I understand it, RUP currently offers no Gold OA option for its three journals, but it is happy for authors to self-archive their papers six months after publication (Green OA). Indeed, RUP itself makes all the papers it publishes feely available six months after publication, on its own web site, in PubMed Central (PMC), and in Europe PubMed Central. As such, it supports and endorses Green OA, and is currently compliant with all the major OA policies, including those of the Wellcome Trust, National Institutes of Health (NIH) and Research Councils UK (RCUK). Is that correct?

MR: That’s correct.

RP: Nevertheless, you have concerns about the new Wellcome Trust and RCUK policies, both of which come into effect on 1st April. Specifically, you worry about their requirement that publishers who offer Gold OA must make the papers they publish available under a CC-BY licence. Since RUP itself offers no Gold option I assume your concern is that, as RCUK expects its funded authors to “prefer” Gold over Green OA, some potential RUP authors may feel they should not submit their papers to RUP? Or is your concern more general than that?

MR: My concern is more general than that. RCUK and Wellcome are fragmenting the licensing of their own content by requiring CC-BY when they pay immediate access fees.

If the publisher does not offer an immediate access option but releases the content at six months, Wellcome Trust leaves publishers’ licensing terms in place, and those terms could be as far to the other side of the openness spectrum as the publisher holding copyright with all rights reserved. RCUK requires content released at six months to be made available “without restriction on non-commercial re-use,” effectively CC-BY-NC.

In my opinion, the middle-ground of CC-BY-NC should be mandated for all content. This licence could be used by any publisher regardless of business model, and, I believe, it still achieves the funding agencies’ goal of reuse of content for text and data mining.

I think RCUK and Wellcome should be using their influence to encourage publishers to apply a more open licence (such as CC-BY-NC) universally, rather than calling for the use of CC-BY for just a small proportion of publishers’ content.

RP: What I also think worries you about the CC-BY licence is that if a subscription publisher were to use it, third-party commercial organisations would be able to bundle and re-publish its papers, thereby undermining its subscription revenues. I suspect that Wellcome and RCUK might respond to this concern by saying that in the age of the Internet scholarly communication should be moving to the position where all research papers are made freely available from the day of publication, not six months, or twelve or more months, later.

And they might add that if a subscription publisher has concerns that their policies will have a negative financial impact on it then all it needs to do is embrace Gold OA and levy an article-processing charge, which they are happy to fund for researchers. Is that not a reasonable position for a funder to take?

MR: As I noted in the editorial, we do not offer an immediate access option because we do not believe in providing immediate access to just a subset of our content. We prefer to release all of our content at six months.

In addition, I believe that APCs for immediate access to a subset of content in an otherwise subscription-based journal [Hybrid OA] can unfairly distribute the costs of publication. As a publisher of “selective” journals, our cost per article is higher than the current standards for APCs for immediate access, even the $5,000 fee that is now charged by some journals.

I believe it sends the wrong message to the research community for a subscription-based journal to make an article free immediately for less than the actual cost of doing so. In such a case the immediate access fee is being subsidized by subscribers, who may not want to subsidize immediate access to a subset of articles. 

RP: How much do you estimate it costs RUP to publish a paper? I.e. what APC level would it need to charge? 

MR: I specified a cost per article of ~$10,000 three years ago. I have not redone the calculation since then. 

No plans to become an OA publisher


RP: Would it be right to conclude that RUP is happy with the subscription model, and has no current plans to become an OA publisher? Does it envisage every giving up its subscription model?

MR: RUP has no current plans to become an OA publisher, although we remain open to the possibility if we could be assured of appropriate support from funding agencies to cover our expenses as a selective journal.

For the time being, I believe that the subscription-based business model for selective journals can and should coexist with the OA business model.

RP: The key point about the new licensing requirement, I guess, is that it is being imposed by RCUK and Wellcome because they believe that research that they have funded should be freely available for third parties — not just to read, but to text and data mine too. And in order for this to be possible, they say, the content will need to be published under a CC-BY licence.

You point out that this is also possible if a publisher licences its papers under a CC-BY-NC-SA — which I think is similar in its “viral effect” to Richard Stallman’s General Public Licence (GPL). And that is what RUP does. Have I understood correctly?

MR: That’s correct, although I can’t comment on the terms of the GPL.

RP: I believe you also argue that text and data mining is already permissible under fair use exemptions. While you agree there is some uncertainty over this, you think a better solution here might be to clarify current fair use rules, not force publishers to adopt a licence that they are not comfortable with. Is that right? And given what you say, do you think that the CC-BY licence requirement being imposed by Wellcome/RCUK is superfluous?

MR: That’s correct. As I noted above, I think the goals of Wellcome and RCUK for open reuse of content can be achieved with a licence that is compatible with a subscription-based business model.

RP: You also say in your editorial that the RCUK and Wellcome licence requirements will initially create considerable confusion over what third parties can do with a particular journal’s content — since some of a publisher’s content will be licensed under CC-BY, some of it under different licences.

And as you say above, this will serve to “fragment” the licensing of the funders’ content. Some might argue that since scholarly communication is going through a period of radical transition some confusion is inevitable in the short-term, but that this should be viewed as a necessary period of uncertainty to a greater end. How would you respond to that?

MR: On this particular point, I do not think the confusion is necessary, even in the short term.

RP: A key point about the CC-BY-NC-SA licence you use, of course, is that it forbids commercial reuse, and it requires that anyone reusing the content must do so under the terms of the original licence. There are, however, those who believe that publicly-funded research ought to be reusable for commercial purposes as well as non-commercial.

My understanding is that the primary reason why the UK government wants CC-BY to become the norm is precisely in order to enable commercial reuse, which it apparently believes will have a direct economic benefit to the UK economy. That certainly seemed to me to be the thrust of what the UK Minister of State for Universities and Science David Willetts said to the Publishers Association last May. And it was Willetts who commissioned the Finch Report, which led to the RCUK policy.

Do you not feel that this is a reasonable position for a government minister to take with regard to publicly-funded research? After all, he is happy to pay publishers for the services they provide. And in fact it is publishers who set the APC rate, not governments. All publishers need do is offer a Gold OA option and charge an article-processing fee is it not?

MR: I think it is a reasonable position for a government minister to take. In the US, we have the Baye-Dole Act of 1980, which allows funding recipients to have ownership of inventions/discoveries made with federal funding. Thus, commercial use of the original invention/discovery described in a scholarly article is already spoken for.

Data/text mining from scholarly articles provides an opportunity for secondary commercial use. As noted, in my opinion, this secondary level of commercial reuse of scholarly content constitutes fair use and is not restricted by the NC clause in the CC-BY-NC licence.

Text and data mining


RP: Just to clarify: RUP is not entirely opposed to text and data mining by third parties for commercial purposes. Specifically, it currently permits commercial entities to mine its papers for internal research purposes, and for indexing in order to provide search services? What it objects to is the wide scale reuse permissions that CC-BY permits. Is that accurate?

MR: That is accurate. Since CC-BY permits wholesale resale of content, it has (as you noted earlier) the potential to undermine subscription sales by the publisher.

RP: I do sometimes think that politicians and funders are inclined to conflate the issue of open data (making the data produced during scientific experiments freely available) with that of text and data mining (i.e. allowing third parties to access the data underlying a paper, and to treat papers as though they were data rather than text, and mine them with computers). I note in his talk to the PA, Willetts said “every dollar of federal investment in the Human Genome Project has helped generate $141 for the US economy.” Do you think that there is confusion here, and that some might in any case have rather inflated expectations about the potential benefits of text and data mining? Can making research papers freely available for reuse really have a direct and meaningful financial impact on national economies in your view?

MR: As you note, access to research data is distinct from access to data underlying a publication, and distinct from access to published articles in machine-readable form. And the rights to mine those data (and text) are another issue.

In last year’s OSTP Request For Information on Public Access to Digital Data Resulting From Federally Funded Scientific Research, I argued that US federal mandates should require access to data underlying a publication and should provide the funds necessary to create the mechanisms to provide such access.

I do think there is potential for secondary discovery through text and data mining, and I support any technical or licensing efforts to enable this type of discovery. I believe that such efforts can be made in a way that is compatible with a subscription-based business model.

RP: Can I ask if you support the US Fair Access to Science and Technology Research Act (FASTR) — which is a new, stronger version of FRPAA? I note Creative Commons has pointed out that while not directly requiring the application of open licences, FASTR “includes provisions that move the ball down the field toward better communicating reuse rights”. Would I be right in thinking that you would rather not see the ball moving in that direction?

MR: I do support FASTR, and I think both aspects — access to research articles and licensing terms for reuse — can be implemented in ways that are not detrimental to subscription-based publishers. As I noted in an earlier Q&A session on your blog about FRPAA:

Although it can be argued that a six-month embargo period may not be suitable for all disciplines covered by FRPAA, this is not grounds to oppose the legislation altogether. It should be supported in principle and could be modified during Congressional review to provide the flexibility for each agency to choose its own embargo period.

Thus, I think FASTR would be more likely to succeed if it were modified to match the approach of the recent White House Directive, which recognizes that individual fields of research may need different embargoes.

In 2001, The Rockefeller University Press decided to release its content to the public six months after publication based on the decay rates of hits on research articles, and I remain convinced that this is a suitable delay for biomedical research. I have not seen similar data for other fields. Have you?

And I do support licensing terms that provide reuse rights. I just don’t think it is necessary to invoke a licence that is potentially detrimental to subscription-based publishers to do so.

I think I proved my support for reuse rights by adopting CC-BY-NC-SA licensing for all of our content in 2008 and by now adding the clarification to our licensing policy indicating that it permits text and data mining by anyone, including commercial entities.

RP: Thank you for taking the time to answer my questions.

7 comments:

  1. Mike

    I'd like to respond to a couple of the points you raise, but first let me say that RUP - by virtue of the fact that it makes all its content freely available at PMC after 6 months - continues to be a "fully compliant" publisher with the Wellcome Trust and RCUK.

    Let me now turn to a couple of issues you raise.

    You make the point that content gets fragmented if we allow both gold and green. I accept this, and though Wellcome Trust policy has always allowed green in certain circumstances (i.e. Nature, Science, NEJM, JAMA etc do not provide a gold option) we make no secret of the fact that we prefer gold. When green is chosen we have never felt we can mandate the way in which such works are licensed but have encouraged – and worked with publishers – to liberalise their policies to the greatest possible extent when green is the only option. E.g. Nature’s policy on TDM, http://www.nature.com/authors/policies/license.html#terms

    You also argue that CC-BY is not necessary for TDM. This may be true under the US copyright doctrine of fair use (although this is largely untested) but it is not the case under the laws of the UK and other jurisdictions.

    Even if it were the case that CC-BY was not required for TDM, an NC restriction would still prevent the widest possible use of the literature by all, as it would rule against use by commercial companies. Wellcome Trust does not fund research for the outputs to be restricted for use only to academics.

    Furthermore, the NC element limits the way in which the research we have funded can be used, as demonstrated by these examples:

    a) Posting content on a third-party site

    Although content will always be freely available on PMC and Europe PMC, it may also be desirable to make that content available on another site. For example, if we funded some research into HIV transmissions, it may be considered desirable to post this information directly on to web sites where communities at most risk from this disease would see it. However, if those site carried any advertising, then that would breach the “non-commercial” clause, and as such it is likely that the article would be subject to a take-down request.

    Equally, the NC clause prohibits a user from sharing content at resources like Wikipedia. See table at: http://en.wikipedia.org/wiki/Wikipedia:FAQ/Copyright)

    b) Creating a translation and then charging for this.

    So, the Trust funds a lot of research into malaria and its management and prevention. It is possible that an organisation may like to take this content and translate it (into, say, Burmese) so the information could be more readily understood by the local population. Creating the translation may incur costs, which the creator may wish to recover by selling this content. As long as the translation attributes the original research (which would remain freely available) this re-use is permissible. With NC such a use would not be permissible

    ReplyDelete
  2. OVER-REACHING

    1. Mike Rossner is right in all but the RUP 6-month embargo on Green OA. RUP should of course allow immediate Green OA. But the OA embargo can be lived with for now: Once immediate-deposit mandates (regardless of whether the deposit is immediately OA or OA is embargoed) prevail worldwide, OA embargoes will die their inevitable natural deaths as a result of the increasingly palpable benefits of OA. All that's needed is the mandates and a little patience and confidence in the benefits of OA. Needlessly continuing instead to over-reach pre-emptively (e.g., for zero embargoes) simply serves to keep delaying the optimal and inevitable outcome (as it has done for 20 years).

    2. The same answer for Robert Kiley: The Welllcome Trust's preference for immediate Gold CC-BY is over-reaching, needlessly, and thereby delaying (embargoing!) the success and worldwide growth of Green OA (immediate-deposit) mandates. Counterproductive license restrictions will die their inevitable natural deaths as a result of universally mandated immediate-deposit Green OA. All that's needed is the mandates and a little patience and confidence of OA. Needlessly continuing instead to over-reach pre-emptively (e.g., for zero-restriction licenses) simply serves to keep delaying the optimal and inevitable outcome (as it has done for 20 years).

    ReplyDelete
  3. Robert Kiley says: Equally, the NC clause prohibits a user from sharing content at resources like Wikipedia. See table at: http://en.wikipedia.org/wiki/Wikipedia:FAQ/Copyright)

    Comment: first, I would like to say that I am a fan of Wikipedia. I use Wikipedia, have contributed to Wikipedia in the past, and would like to see more scholarly content in Wikipedia. However, there are some problems with including scholarly content in Wikipedia, and I argue that these are not resolved by CC-BY.

    The first problem is the difference in the expectations of attribution. For scholars, the key elements in attribution are things like the scholar, journal, and publisher. In Wikipedia, the norm is Attribution. This means that Wikipedia and scholarly works are not compatible from the perspective of Attribution.

    Another reason for caution in including scholarly works in Wikipedia is the norm that anyone can edit. If the work of a serious scholar is deposited in Wikipedia, it could be "corrected" by someone with far less knowledge. If this is combined with Attribution (assuming the Wikipedia anonymity problem is overcome), then it seems highly likely that the result will be incorrect citations of scholar's works. This would harm rather than benefit scholarship.

    Finally, if we want to see more scholarly works in Wikipedia, it is Wikipedia policies and practices that need to change to accomodate this rather than a wholesale transformation of scholarship to accomodate the preferences of the current Wikipedia team.

    To return to Robert Kiley's comment: itis not the NC clause that prohibits a user from sharing content at Wikipedia, but rather Wikipedia's policies that do not permit works with NC clauses. Wikipedia is free to develop a more flexible policy at any time.

    For more on my analysis of the imperfect match of Creative Commons and Open Access see my Creative Commons and open access critique series.

    ReplyDelete
  4. Mike Rossner has asked me to post this below in response to Robert Kiley's comments. I have had to split the text into two sections due to the limitations of Blogger’s comment functionality.

    PART ONE

    RK: You make the point that content gets fragmented if we allow both gold and green. I accept this, and though Wellcome Trust policy has always allowed green in certain circumstances (i.e. Nature, Science, NEJM, JAMA etc. do not provide a gold option) we make no secret of the fact that we prefer gold. When green is chosen we have never felt we can mandate the way in which such works are licensed but have encouraged – and worked with publishers – to liberalise their policies to the greatest possible extent when green is the only option. E.g. Nature’s policy on TDM.

    MR: I appreciate that Wellcome's ideal is to have all articles resulting from its funding to be available on an immediate access basis; however, I continue to think that you are not using your influence to its fullest extent to move publishers toward making all of their content available to the public under more open licenses. The liberalized policy at Nature to which you refer applies to only "Articles published by Nature Publishing Group (NPG) which are made available [by someone else, not by NPG] through academic repositories" - another example of fragmentation of policies. I think there is an opportunity and a need to push for more global changes in policy.

    RK: You also argue that CC-BY is not necessary for TDM. This may be true under the US copyright doctrine of fair use (although this is largely untested) but it is not the case under the laws of the UK and other jurisdictions.

    MR: How does Google index content from the UK?

    ReplyDelete
  5. PART TWO

    RK: Even if it were the case that CC-BY was not required for TDM, an NC restriction would still prevent the widest possible use of the literature by all, as it would rule against use by commercial companies. Wellcome Trust does not fund research for the outputs to be restricted for use only to academics.

    Furthermore, the NC element limits the way in which the research we have funded can be used, as demonstrated by these examples:

    a) Posting content on a third-party site.

    Although content will always be freely available on PMC and Europe PMC, it may also be desirable to make that content available on another site. For example, if we funded some research into HIV transmissions, it may be considered desirable to post this information directly on to web sites where communities at most risk from this disease would see it. However, if those site carried any advertising, then that would breach the “non-commercial” clause, and as such it is likely that the article would be subject to a take-down request.

    Equally, the NC clause prohibits a user from sharing content at resources like Wikipedia. See table.

    MR: I understand that the term "non-commercial" is open to interpretation and its definition may vary between countries. At least in the US, not all websites that accept advertising revenue are considered commercial - our own websites at The Rockefeller University Press (a non-commercial entity) are such an example. Following this definition, RUP would not block third-party hosting in the use case you describe. What we aim to prevent with the NC restriction is commercial reuse in the form of resale of the content (or use of the content for direct promotion of a commercial entity) without permission.

    RK: b) Creating a translation and then charging for this.

    So, the Trust funds a lot of research into malaria and its management and prevention. It is possible that an organisation may like to take this content and translate it (into, say, Burmese) so the information could be more readily understood by the local population. Creating the translation may incur costs, which the creator may wish to recover by selling this content. As long as the translation attributes the original research (which would remain freely available) this re-use is permissible. With NC such a use would not be permissible.

    MR: Sale of a translation would be permissible under NC with the permission of the copyright holder or licensee, who may prefer this route because it could provide the opportunity to monitor the accuracy of such a translation.

    ReplyDelete
  6. Heather, above, writes "Wikipedia is free to develop a more flexible policy at any time." This seems to be a misconception. Wikipedia is "the free encyclopedia" (meaning libre) and its eleven sister projects are similarly free. It is absolutely core to what Wikipedia *is* that its content is freely reusable by anyone. The concept of "free" that is used includes commercial use, for good reasons. Introducing non-commercial clauses would block a great deal of legitimate dissemination of Wikipedia and prevent its stated goal of making the sum of human knowledge freely available to everyone on the planet.

    Wellcome's and other funders' priority is understandably to create maximum possibilities for innovative reuse: that means liberating the content. So Wikipedia isn't the core issue- free content is. The idea that Wikipedia can give up its commitment to free content (it can't) is a red herring.

    As for the scenario of "incorrect citation of scholar's works": this needs to be spelled out. Apart from some of the top professional publishers, who is more concerned about correct citation than the Wikipedia community? The possibility that edits might be made to Wikipedia is no reason for a funder not to want the best quality academic text and media to be compatible with it. Editing is evolutionary: we're all here because of a process where *random* changes could happen to a collection of information.

    Stevan accuses Wellcome Trust of overreaching. He will be vindicated if researchers don't take up Wellcome funds because they don't accept the gold OA requirement.

    Thanks Robert for clearly setting out the commitment to free content.

    ReplyDelete
  7. Documents of any kind released under Non Commercial use only are useless for anyone serious about research dependent on international data-mining.

    As "fair use" cannot be argued across the countries of Europe, any research suddenly becomes far too expensive, in terms of the effort required to check individual copyright releases and establish agreements for each source. Sources such as RUP will invariably get ignored in global searches, not a good position for academics choosing where they would like to be published in order to get noticed and cited elsewhere.

    It is surprising that RUP's strategy seems entirely governed by a narrow provincial perspective.

    ReplyDelete