How to open a scientific journal
We started an ambitious project - to open our electronic scientific journal. At first it seemed that this business was not very heavy and nothing good would come of it, especially since we had never been engaged in publishing. However, as with any business, the main thing is to start. Although the future of our journal is still under question, but I decided to describe our experience on this difficult path and, I hope, this story will encourage someone else to create their own good journals for the benefit of Russian science.
We wanted something like this: to create an electronic peer-reviewed journal in English, completely official, which would be taken seriously by Western scientists, to which articles would be used to calculate the impact factor . The minimum program is to get on the list of VAK journals , ideally, to get to PubMed (our journal on bioinformatics). Commercial gain was not expected.
The first thing, of course, is to install some kind of web application for the magazine. Theoretically, you can publish articles even in WordPress, but still the magazines have their own specifics. We need support for issues, sections, review of articles and much more, which became clear along the way. At first glance, there is quite a lot of free special software for logging. Here, for example, a list of as many as 13 points. Here is a review of the software. We had a chance to try two options.
This system attracted us primarily to several things. Firstly, it is called with the letter A, because it comes first in the above list (marketers, use us!). Secondly, it is officially used for all journals on the biological topics of the PLOS family, which means that the specifics of the subject area can be better taken into account in this system. Finally, Ambra is written in Java, under Tomcat, and these technologies are familiar to us.
')
Unfortunately, the benefits have ended here. It seems that PLOS writes it mainly for itself and, although it spreads the source code, it’s hard work to deploy and configure everything to strangers. The installer is not included, from the useful documentation only Quick-start on the wiki (which is not so quick), and what to do next is unclear. Constantly it seems that the heaps of the necessary files are simply not enough, or something needs to be pumped out of the repository and put somewhere, but it is not clear what and where. There is a half-dead mailing list, which does not help much. After tormenting a few days with Ambra, I threw it and began to search further.
This is a PHP application that left a much better impression, and in the end we stopped at it. OJS boasts a much more impressive list of magazines that use it. It was installed easily (most importantly, do not mistakenly download the developer version, which for some reason on the Download page is more stable). The system allows you to host multiple logs at once; supports access rights, subscriptions, and roles; integrates with PayPal to get money from subscribers; allows you to publish an article in several formats (HTML, PDF) and most importantly - it has the tools to clean up the references (unfortunately, this component has some unpleasant bugs). New useful features were regularly found. The code is intelligible and structured, so if you don’t like something, you can easily tweak it. There are quite lively forum, bugtracker and wiki. Among the minuses, in addition to the buggy bibliographic tools, we can mention incomplete and slightly outdated official documentation .
The format for storing and submitting articles should be decided immediately. OJS does not impose any particular format, it only offers to download, for example, ready-made HTML and PDF versions of articles. It is assumed that the layout designer will prepare them himself in third-party applications. For many, LaTeX comes to mind as an internal format that is easy to convert to both HTML and PDF. To translate into LaTeX, however, it’s up to the typesetter himself: if for a mathematical journal it’s still possible to force all authors to send articles to LaTeX, in biology it’s an empty number. Moreover, in this format, then both reviewers and editors will have to work. One way or another, but you have to accept the fact that the articles will come in Microsoft Word formats.
However, as it turned out, LaTeX is still not the best option. In order to get to PubMed, you must first get to PubMedCentral, and this requires that all articles be in the XML of the special scheme of the National Medical Library (NLM Journal Publishing Tag Set), or in another widely used scheme. It is quite reasonable articles to this type and lead. There are two problems: how to automate the creation of this XML and how to automate the generation of HTML and PDF from it.
The generation of the list of references in NLM formats is already supported in OJS, which is very pleased! To generate blanks with the name, annotation, all necessary output data and a list of references, I wrote a special template in OJS. Fully automatically convert docx file to NLM will not succeed. Even if you create a special template for MS Word, anyway, the authors will not prepare articles strictly on it. No, no, yes, the title will be introduced not by style, but simply by increasing the size. Swearing with authors is longer than fixing it manually, and even starting a journal in general is unprofitable to swear with authors. Therefore, we must still create a template for MS Word with the requirements, automate the maximum of what is possible, and manually finish the rest.
We unpack the docx file (still aware that this is a zip archive?) Document.xml and perform a custom XSL transform that maximally converts its contents into what is needed. To convert use Saxon . After that, some trivia dopinivaem Perl-script, then insert the result in a blank and right hands.
To generate HTML from NLM, there is a ready-made XSL transformation on the NLM site . It, of course, had to be modified a little to satisfy our ideas of style, but on the whole there were no problems. It turned out to be more difficult with the generation of PDF. For the same link, there is a conversion to the XSL-FO format, but the free Apache FOP FO processor refuses to chew the result, cursing the use of unsupported features. Contacting paid FO processors like the Antenna House Formatter is a bit reluctant. There was a thought to generate PDF from HTML. From a separate software that solves this problem, we tried wkhtmltopdf , which uses the WebKit engine for rendering HTML. Unfortunately, it turned out that it does not support formulas in MathML . As a result, at the moment everything came down to using the Firefox + PDF Creator bundle. The generated HTML adds the commands window.print () and window.close (); in a separate Firefox installation, quick printing is allowed without a dialog with parameters and allowed web pages to close their window. For better PDF pagination, the page-break-inside CSS property is used in places. Unfortunately, the generated PDF is far from ideal. What is particularly lacking is a two-column layout with the removal of pictures and tables from the general flow and gluing them to the top and bottom edges of the page. In general, there is something to improve. Other PDF magazines are hand-crafted by special people, but we do not yet have the resources.
About the formula should be said specifically. NLM XML uses MathML to represent formulas. In the files that the authors send, the formulas can be seen in the form of Microsoft Equation objects, MathType , in the form of native formulas that appeared in Microsoft Word 2007 or in general in the form of pictures. Well, we just banned sending the pictures, and everything else can be converted to MathML using the same MathType. When generating the HTML version of the article, we leave the formulas as they are, just remove the namespace mml. MathML is poorly supported by browsers, but using MathJax saves. When generating a PDF, the native MathML renderer Firefox is used.
In order for people to know about the created magazine, articles from it were located in special search engines and could normally be referenced to it, you need to register in a bunch of places. Here, too, come across all sorts of difficulties.
Getting an International Standard Serial Number is perhaps the first and simplest thing to do. Logs without ISSN elsewhere may not be considered. Receiving ISSN for free and for electronic journals everything can be done online . Fill out the form, send, wait 10 working days, you get a number. Obtaining an ISSN can and should be done before the publication of the first issue of the journal (but not earlier than six months). You will be given a number that will not be officially registered in the ISSN registry. After the first issue of the magazine you provide a link to it, and your number is made official.
DOI - digital object identifier; A stable identifier that is assigned to journals, issues, articles, and other materials published online. It looks, for example, like this: 10.1126 / science.338.6114.1558 . DOI does not change even when moving the magazine to another publisher, therefore, referring to an article using DOI, you can be sure that the link will remain alive. Without the DOI, the scientific journal is nowhere, but you have to pay decent for them. There are several DOI registrars, but the competition is not as strong as, for example, with domain name registrars. Actually, besides the CrossRef of other registrars it is difficult to consider seriously. CrossRef takes at least $ 275 a year from the publisher , plus a dollar for each new DOI allocated for an article or magazine issue. To register there you need to view the hour webinar and send a contract to the States by courier. If everything works out, you will be sent the first invoice and after its payment you will receive a DOI-prefix. The main requirements are to publish a unique material (without copy-paste) and in the references in the published articles indicate the DOI-links.
OJS supports DOI quite well: you can specify an arbitrary format for generating new identifiers, OJS automatically assigns them to new articles and generates XML for registering new DOIs with CrossRef.
WorldCat is a unified directory of tens of thousands of libraries around the world. They support the OAIster service, which can index your articles. Register there for free and you can before the first issue of the magazine. WorldCat will independently download and index meta information about articles in the format of the Open Archives Initiative (OJS supports this format).
To participate in the Russian Science Citation Index ( RISC , this is useful for getting into the list of HAC), you will need to conclude an agreement with eLibrary.ru . This is free, but it has added additional complexity: they need to provide metadata (authors' names, their organizations, titles and annotations of articles) in Russian and English, even if the article is published only in English (or only in Russian).
Despite all the shortcomings, the impact factor remains the main criterion for the quality of a scientific journal. To get into the system of calculating the impact factor, you must register the journal in the Web of Science . There is already estimated the scientific component. To review the journal, you must submit three consecutive issues, and you must submit an application after the release of the first, and then again after the second and third. If everything goes well, in two years you will be assigned an impact factor.
PubMedCentral (PMC) is a free full-text archive of articles on biology and related topics. Almost all Western biological journals publish their articles in PMC. In addition, it is the most direct route to PubMed. In general, any journal should consider the possibility of cooperation with any scientific archive (besides PMC there are, for example, CLOCKSS or Portico ). This ensures that the articles are not lost to the scientific world, even if your publishing house ceases to exist.
After submitting the application, PMC first of all evaluates the scientific component of the journal and, if successful, evaluates the technical side. There are quite strict requirements. As I said above, articles should be in XML of a particular schema, and PMC imposes additional requirements . Special rules exist for images in articles . If everything went well, then a seven-page contract is concluded, which must be sent to PMC.
If you have overcome all the technical difficulties, then it remains just a little - to find authors, reviewers, editors and layout designers who want to work with your journal. How and where to look for them - these questions are beyond the scope of my article.
We wanted something like this: to create an electronic peer-reviewed journal in English, completely official, which would be taken seriously by Western scientists, to which articles would be used to calculate the impact factor . The minimum program is to get on the list of VAK journals , ideally, to get to PubMed (our journal on bioinformatics). Commercial gain was not expected.
Soft
The first thing, of course, is to install some kind of web application for the magazine. Theoretically, you can publish articles even in WordPress, but still the magazines have their own specifics. We need support for issues, sections, review of articles and much more, which became clear along the way. At first glance, there is quite a lot of free special software for logging. Here, for example, a list of as many as 13 points. Here is a review of the software. We had a chance to try two options.
Ambra project
This system attracted us primarily to several things. Firstly, it is called with the letter A, because it comes first in the above list (marketers, use us!). Secondly, it is officially used for all journals on the biological topics of the PLOS family, which means that the specifics of the subject area can be better taken into account in this system. Finally, Ambra is written in Java, under Tomcat, and these technologies are familiar to us.')
Unfortunately, the benefits have ended here. It seems that PLOS writes it mainly for itself and, although it spreads the source code, it’s hard work to deploy and configure everything to strangers. The installer is not included, from the useful documentation only Quick-start on the wiki (which is not so quick), and what to do next is unclear. Constantly it seems that the heaps of the necessary files are simply not enough, or something needs to be pumped out of the repository and put somewhere, but it is not clear what and where. There is a half-dead mailing list, which does not help much. After tormenting a few days with Ambra, I threw it and began to search further.
Open Journal Systems (OJS)
This is a PHP application that left a much better impression, and in the end we stopped at it. OJS boasts a much more impressive list of magazines that use it. It was installed easily (most importantly, do not mistakenly download the developer version, which for some reason on the Download page is more stable). The system allows you to host multiple logs at once; supports access rights, subscriptions, and roles; integrates with PayPal to get money from subscribers; allows you to publish an article in several formats (HTML, PDF) and most importantly - it has the tools to clean up the references (unfortunately, this component has some unpleasant bugs). New useful features were regularly found. The code is intelligible and structured, so if you don’t like something, you can easily tweak it. There are quite lively forum, bugtracker and wiki. Among the minuses, in addition to the buggy bibliographic tools, we can mention incomplete and slightly outdated official documentation .Layout and article formats
The format for storing and submitting articles should be decided immediately. OJS does not impose any particular format, it only offers to download, for example, ready-made HTML and PDF versions of articles. It is assumed that the layout designer will prepare them himself in third-party applications. For many, LaTeX comes to mind as an internal format that is easy to convert to both HTML and PDF. To translate into LaTeX, however, it’s up to the typesetter himself: if for a mathematical journal it’s still possible to force all authors to send articles to LaTeX, in biology it’s an empty number. Moreover, in this format, then both reviewers and editors will have to work. One way or another, but you have to accept the fact that the articles will come in Microsoft Word formats.
However, as it turned out, LaTeX is still not the best option. In order to get to PubMed, you must first get to PubMedCentral, and this requires that all articles be in the XML of the special scheme of the National Medical Library (NLM Journal Publishing Tag Set), or in another widely used scheme. It is quite reasonable articles to this type and lead. There are two problems: how to automate the creation of this XML and how to automate the generation of HTML and PDF from it.
The generation of the list of references in NLM formats is already supported in OJS, which is very pleased! To generate blanks with the name, annotation, all necessary output data and a list of references, I wrote a special template in OJS. Fully automatically convert docx file to NLM will not succeed. Even if you create a special template for MS Word, anyway, the authors will not prepare articles strictly on it. No, no, yes, the title will be introduced not by style, but simply by increasing the size. Swearing with authors is longer than fixing it manually, and even starting a journal in general is unprofitable to swear with authors. Therefore, we must still create a template for MS Word with the requirements, automate the maximum of what is possible, and manually finish the rest.
We unpack the docx file (still aware that this is a zip archive?) Document.xml and perform a custom XSL transform that maximally converts its contents into what is needed. To convert use Saxon . After that, some trivia dopinivaem Perl-script, then insert the result in a blank and right hands.
To generate HTML from NLM, there is a ready-made XSL transformation on the NLM site . It, of course, had to be modified a little to satisfy our ideas of style, but on the whole there were no problems. It turned out to be more difficult with the generation of PDF. For the same link, there is a conversion to the XSL-FO format, but the free Apache FOP FO processor refuses to chew the result, cursing the use of unsupported features. Contacting paid FO processors like the Antenna House Formatter is a bit reluctant. There was a thought to generate PDF from HTML. From a separate software that solves this problem, we tried wkhtmltopdf , which uses the WebKit engine for rendering HTML. Unfortunately, it turned out that it does not support formulas in MathML . As a result, at the moment everything came down to using the Firefox + PDF Creator bundle. The generated HTML adds the commands window.print () and window.close (); in a separate Firefox installation, quick printing is allowed without a dialog with parameters and allowed web pages to close their window. For better PDF pagination, the page-break-inside CSS property is used in places. Unfortunately, the generated PDF is far from ideal. What is particularly lacking is a two-column layout with the removal of pictures and tables from the general flow and gluing them to the top and bottom edges of the page. In general, there is something to improve. Other PDF magazines are hand-crafted by special people, but we do not yet have the resources.
About the formula should be said specifically. NLM XML uses MathML to represent formulas. In the files that the authors send, the formulas can be seen in the form of Microsoft Equation objects, MathType , in the form of native formulas that appeared in Microsoft Word 2007 or in general in the form of pictures. Well, we just banned sending the pictures, and everything else can be converted to MathML using the same MathType. When generating the HTML version of the article, we leave the formulas as they are, just remove the namespace mml. MathML is poorly supported by browsers, but using MathJax saves. When generating a PDF, the native MathML renderer Firefox is used.
check in
In order for people to know about the created magazine, articles from it were located in special search engines and could normally be referenced to it, you need to register in a bunch of places. Here, too, come across all sorts of difficulties.
ISSN
Getting an International Standard Serial Number is perhaps the first and simplest thing to do. Logs without ISSN elsewhere may not be considered. Receiving ISSN for free and for electronic journals everything can be done online . Fill out the form, send, wait 10 working days, you get a number. Obtaining an ISSN can and should be done before the publication of the first issue of the journal (but not earlier than six months). You will be given a number that will not be officially registered in the ISSN registry. After the first issue of the magazine you provide a link to it, and your number is made official.DOI
DOI - digital object identifier; A stable identifier that is assigned to journals, issues, articles, and other materials published online. It looks, for example, like this: 10.1126 / science.338.6114.1558 . DOI does not change even when moving the magazine to another publisher, therefore, referring to an article using DOI, you can be sure that the link will remain alive. Without the DOI, the scientific journal is nowhere, but you have to pay decent for them. There are several DOI registrars, but the competition is not as strong as, for example, with domain name registrars. Actually, besides the CrossRef of other registrars it is difficult to consider seriously. CrossRef takes at least $ 275 a year from the publisher , plus a dollar for each new DOI allocated for an article or magazine issue. To register there you need to view the hour webinar and send a contract to the States by courier. If everything works out, you will be sent the first invoice and after its payment you will receive a DOI-prefix. The main requirements are to publish a unique material (without copy-paste) and in the references in the published articles indicate the DOI-links.OJS supports DOI quite well: you can specify an arbitrary format for generating new identifiers, OJS automatically assigns them to new articles and generates XML for registering new DOIs with CrossRef.
Worldcat
WorldCat is a unified directory of tens of thousands of libraries around the world. They support the OAIster service, which can index your articles. Register there for free and you can before the first issue of the magazine. WorldCat will independently download and index meta information about articles in the format of the Open Archives Initiative (OJS supports this format).eLibrary.ru
To participate in the Russian Science Citation Index ( RISC , this is useful for getting into the list of HAC), you will need to conclude an agreement with eLibrary.ru . This is free, but it has added additional complexity: they need to provide metadata (authors' names, their organizations, titles and annotations of articles) in Russian and English, even if the article is published only in English (or only in Russian).Impact factor
Despite all the shortcomings, the impact factor remains the main criterion for the quality of a scientific journal. To get into the system of calculating the impact factor, you must register the journal in the Web of Science . There is already estimated the scientific component. To review the journal, you must submit three consecutive issues, and you must submit an application after the release of the first, and then again after the second and third. If everything goes well, in two years you will be assigned an impact factor.
PubMedCentral
PubMedCentral (PMC) is a free full-text archive of articles on biology and related topics. Almost all Western biological journals publish their articles in PMC. In addition, it is the most direct route to PubMed. In general, any journal should consider the possibility of cooperation with any scientific archive (besides PMC there are, for example, CLOCKSS or Portico ). This ensures that the articles are not lost to the scientific world, even if your publishing house ceases to exist.After submitting the application, PMC first of all evaluates the scientific component of the journal and, if successful, evaluates the technical side. There are quite strict requirements. As I said above, articles should be in XML of a particular schema, and PMC imposes additional requirements . Special rules exist for images in articles . If everything went well, then a seven-page contract is concluded, which must be sent to PMC.
All the rest
If you have overcome all the technical difficulties, then it remains just a little - to find authors, reviewers, editors and layout designers who want to work with your journal. How and where to look for them - these questions are beyond the scope of my article.
Source: https://habr.com/ru/post/168903/
All Articles