LATEX source from word processors
Contents
1. quoted from The TeX book Writer2LaTeX will aim to produce excellent result for this kind of documents including of course shorter texts with a book like layout This quotation is fair but I think it makes my point go ahead and inhale Show the logos TEX and I4TEX correctly use the correct dashes and spacing use the proper quotes 6 1 Examples of other problems I present an annotated list of a few other problems I addressed in the macros These are based on two of the test cases the rotordynamics text and the programming text I think it is fair to classify most of these as not very clean LTRX Export of spacing The export of chapter 5 of the rotordynamics text has 47 occurrences of a space preceding a right brace The majority of those are in constructs like textit word while most of the rest are weird constructs like textit and textbf textit The first may be sloppy keyboarding by the author The second seems to be intentional Bart Childs TUGboat Volume 33 2012 No 2 spacing why not The last is likely a hacked indentation kludge Inline mathematics Some inline mathematics is converted to italics That is troublesome to me because it should really remain as unconverted mathematics Then too that may be the fault of the author Export of structure The structure of the chap ter and lists range from inconsistent to missing This is likely the authors fau2. The ATEX output is cleaner as most of the reduction in the size is the elimination of need less formatting like 1 most paragraphs were inside IATEX source from word processors 182 grouping braces and a declaration that I used En glish and 2 textquotedblright for a simple A cursory look at the user s guide indicates some removal of redundancy There is a lack of the com pleteness that is characteristic of the documentation of releases from the TEX communities I plan to work with OOo and continue to make this product better I believe it to be the best hope I know of especially in the open domain The following quote is from one of the pages on its web site http writer2latex sourceforge net index3 htm1 You will never get a result that looks identical to the original in fact that s the whole point LaTeX is in general a superior typesetting engine compared to Writer For example LaTeX produces much better results for formulas it has an excellent paragraph and page breaking mechanism it uses ligatures etc On the other hand Writer has a few features that LaTeX does not support well If the layout of your document depends on text flowing around pictures or linked text boxes you will never get good results with Writer2LaTeX According to TeX s author Donald E Knuth TeX is a typesetting system intended for the creation of beautiful books and especially for books that contain a lot of mathematics
3. 178 TEX source from word processors Bart Childs Abstract Hennings CTAN survey is a good starting point when considering projects implied by the title of this article I found it a fair view of most related packages He suggests having one of two goals con verting the document structure or converting the appearance My goal is neither of these I want to produce FATRX source that is accurate in content clean and therefore maintainable This is in keeping with Knuth s original goals in producing TEX graphic excellence and a document convenient for archiving Structure and appearance are important I believe clean ATEX is more likely to have this intrinsic result not use of word processing systems My current conversion system is a hybrid based on the use of the Open Office Writer package its Writer2LaTeX application and macros for the Emacs editor written in Elisp The test cases for this system are books 1 on rotordynamics 2 a C programming text 3 a memoir on a friend s life including significant text fragments in the Czech language and 4 a novel that includes three love triangles Even the worst case with significant math ematics formatting done in WordPerfect is tractable I did not say easy The lack of intelligent use of word processors causes many of the problems I esti mate that a 300 page novel written in a reasonable dialect of Word WordPerfect or Writer could be converted to ATEX in an hour or
4. ave used show that the graphical in terfaces are not a solution to the problems in my humble opinion All the authors are highly educated and familiar with the problems of getting people to learn at the college level Still each has shown the results from casual learning about their systems The effective use of styles consistent use of symbols and special functions document structure etc were lacking in each of their documents TUGboat Volume 33 2012 No 2 The first line of a ATEX document requires a statement of the class of the document There is a finite number of them It does not seem to enter the stream of consciousness for many that if they learned how to type Mary had a little lamb on a machine that there should be at least a small change in the start of a letter to a sweetheart a grocery list or any other class of documents In a moment of frustration I lamented Users avoid using TEX because you have to learn how to do some things while users of Word believe if it takes any non obvious effort to do something it should not be done I raised the question earlier about why educated users of computers seem to get so little from user s guides and manuals Maybe the manuals are poor non existent or not convenient Maybe the easy to use graphics interfaces overwhelmed the authors Maybe these interfaces do not encourage users like these to seek the information they need Maybe they just do not care Wa
5. boarding original documents This list contains three cases where it is more efficient to use text with encoding exports than the converted exports assuming the goal of clean ATEX These came from the rotordynamics text the pro gramming text and the User s Manual These are Tables Tables are exported with all formatting on every cell The usual TEX procedure is to give default formatting in a template and excep tional formatting when needed in a cell Mathematics Text is often used for explanatory purposes in equations Programs as well as verbatim text need special handling Portions of some documents are easier to convert by exporting as text with encoding and then insert ing the formatting by editing Two examples are mathematics that does not convert and formatted code fragments in a processor where font changes are done manually rather than using a package like listings The macros were implemented using the mouse or similarly functioning device to point or highlight in conjunction with function keys In Emacs one can also highlight a region of text by setting the mark and moving the point The function keys can also be modified by use of shift control and alt 5 1 Applying fonts to text In this paragraph there are single words and a three word sequence that are emphasized by changing fonts The default font is changed to italics or typewriter Source exported as text with encoding will have for matting remov
6. ed A similar situation occurs when text is inserted into mathematics code TUGboat Volume 33 2012 No 2 The user can highlight a phrase or click within the single word Then the user presses the appropri ate function key for the formatting command to be inserted with grouping of the appropriate text If the user has clicked within a word then the extent of the word is determined by whitespace delimiters Clicking on whitespace is a special form of this the commands are inserted and the cursor placed on the right brace for user input Instead of highlighting a region the user can use the Emacs form of setting the mark and moving the cursor to the other end of the region I imple mented these functions for bold italic sans serif and typewriter fonts I did not insert the italic correction but easily could have paying attention to the following character I did not because in many cases it is just not needed and besides the user should have some responsibilities The same func tions are reused for simple grouping and the text commands which were used mostly in math modes 5 2 Inline mathematics Inline mathematics is common in the rotordynamics text Most of the resulting mathematics is usually a fraction of a line in length The implementation is like the font changes in the previous subsection A significant difference is that the export processes handling WordPerfect mathematics yields significant artifacts of exces
7. ematics content I expected to pur sue a PhD with a screwdriver approach I was willing to do this based on working from the Word rtf Rich Text File format total extraction of text without formatting and or a mostly automatic conversion that needed tweaking my pipe dream A few years after retirement a friend and col league in the college of engineering asked me for help finding someone to keyboard a new text he was writ ing based on a few dozen of his research papers and related studies The topic of the text is rotordynam ics from small pumps and turbines to large ones as in the main engine of the space shuttle I resurrected my plan and we agreed on the plan of work The draft source of this rotordynamics text is being done in WordPerfect the formatter the author has used for many years Most of the text is being adapted from the author s contributions in the sub ject The current version is approximately 400 pages in length with another 25 to be added The lists of contents figures and tables will likely occupy 18 pages There are hundreds of equations with one of them being a full page 2 The process evolves I started this conversion using the process I had prototyped for the programming text The rotordy namics text was quite a different document because of the large fraction of displayed equations The displayed equations and figures in the rotordynamics text require approximately the same fraction o
8. esponding with a publisher in the Czech Republic about his memoir When he wrote to the publishers and sent it by email the formatting was lost I suggested learning a bit of ATEX converting it to pdf and emailing that He had sent me a draft 179 of the book so I could create some examples The published version 1 was done while I was creating this system Of course I was naive and would still have been so had I not read Horak s note 3 But while waiting I thought I could polish my Emacs macros to handle his Czech problems It was fairly easy and with the improvements in the Writer export process it was really easy I mention this project because it shows evidence of real problems with similar projects That will be discussed later In the abstract I mentioned a novel about three love triangles That project was technically triv ial but also contains the same real problems with conversion of word processor sources 4 Real problems There are several sources of problems that impeded progress in these projects Some of these sources could be avoided by user learning while others resulted from differences in the design and implemen tation of the systems they used The authors had several kinds of problems that automatic conversion did not handle 1 Inconsistent use of functionality 2 Wrong use of functionality 3 Not using available functionality 4 Oops Operator operand placement Misunder standings Myst
9. f space required by figures programs and code fragments in the programming text Most maybe all of the code fragments programs and figures in the pro gramming text were restricted from floating There had to be some manual floats I did some small portions of the rotordynamics book as manual conversions for test cases Some of the equations were manually entered because conver sion of mathematics among word processing systems was generally accepted to be non existent I think that is improving The manual process was based on a having a pdf of the document b editing the rtf file c editing a text file exported from a word TUGboat Volume 33 2012 No 2 processor with some encoding and or d a form of ATEX exported from one of several systems I was delivering TFX source faster than I could have keyboarded it from good copy Still it was unsatis factory because it was mostly a manual process The source documents were done in WordPerfect on a PC and I was doing ATEX on a Mac There are good TeX and Emacs systems for the Mac using Mac OS X Some Emacs systems were not acceptable to me because my system uses function keys I continued to strive for big improvements be cause keyboarding mathematics would be slow A significant improvement came by changing the for mat in which sources were delivered to me The source was 1 edited to remove the graphics from the WordPerfect source and 2 exported in rtf form w
10. icism about style files This quote is in section 1 2 of the Writer2LaTeX User s Manual 4 You can use TEX as a typesetting engine for your OOo documents Writer2LaTeX can be configured to create a ATEX document with as much formatting as possible preserved Note that the resulting ATEX source will be readable but not very clean You will find that Writer2LaTeX uses the principle garbage in garbage out Each of the above examples of garbage in garbage out was present in at least two of the test cases cited Garbage in garbage out may be a bit strong of a description for these but the message is clear For example in the Czech memoir it was certainly appro priate to attempt to show correct accents Horak 3 would be proud It overwhelmed the author s limits of skill with the systems he was using Each of the authors has a doctorate and has taught at major universities They are consistent users of computers but obviously are not the most persistent readers of the formatter manuals Maybe the manuals are poor non existent or not conve nient Maybe the easy to use graphics interfaces overwhelmed the authors Maybe these interfaces IATEX source from word processors 180 do not encourage users like these to seek the infor mation they need Maybe they just do not care 4 1 Inconsistent use of functionality The author of the memoir that used many Czech words phrases and sentences is to be saluted fo
11. ith 3 the graphics elements put in a zip file The version of WordPerfect being used would create rtf files hundreds of times bigger than needed if the graphics was included in the export to the rtf Re moving the graphics was no loss because it like the mathematics was not being exported I would take the rtf from WordPerfect im port it to OOo Writer and save it This apparently lost nothing but gave a smaller file and therefore my system was faster in using it I also noticed that Writer s export of text with encoding was different from the other systems I had used Further the ex port could be done in Unicode which was compatible with Emacs Apparently there was significant appreciation of Unicode in the WordPerfect export process The export of the mathematics from WordPerfect was not converted but many symbols Greek letters etc were now viewable on the screen Most IA TpX users should be able to glean the proper content from a printed pdf of the WordPerfect Now the Emacs macros could do much more At this time my benefactor had other obligations and so I had time to work on the macros and test the system using the modified process I continued to learn more Elisp 3 Keeping the mind busy My benefactor s diversions lasted a bit longer than planned I read more about Unicode and realized how provincial some of us are here in the English only USA A college buddy of mine is a Czech immigrant and was corr
12. lt as the use of styles seems to be the cause 7 Writer and friends In spite of these remarks I salute OOo I believe that the Writer package and Writer2LaTeX appli cation have made a great contribution to the goal of converting many documents into a form for better presentation and archival namely ATEX That may not have been the intent The intent may have been to enable a good Writer user to simply use PTEX as an output device The ATEX code output in version beta 1 2 is improved but not clean The Writer2LaTeX User s Manual is 45 pages in length The exported TEX with the clean option source averages about four teen occurrences of mdseries and twelve occur rences of textstyleSourceText per page Each paragraph is grouped with mdseries as the start The latter is effectively an alias for texttt and used in tables 8 Conclusions Reasonable document interchange and archival qual ity is now possible for a wide range of systems I believe that I4 TEX is the most reasonable basis for many archival systems The advances by OOo and its Writer system are impressive and appreciated I hope that its open status and development will continue Note I have addressed only a small part of a large project OOo A point made in a number of venues is the prob lem of TEX systems not having a native graphical input process Lyx and OOo are touted as solutions along with several others The authors of the three test cases I h
13. r attempting to make that text look proper to a Czech reader There are five special items in this sentence On my next visit to Prague he joined Vl a and me along with our wives for lunch at a French restaurant in Obecn dim Municipal House The nickname Vlad a has an accent over the letter a and an accent often called a caron modifying the letter d The accented i in the first italicized word is a dotless i Finally the second italicized word has an accent that almost appears to be the degrees as in temperature symbol Although it was not the author s intention the distances these accents were raised or kerned differed in most cases I do not claim my caron here is perfect 4 2 Misuse of functionality In the rotordynamics book there were many instances of using different Greek characters as the same the phi and varphi and y as well as others Since this document was constructed using papers written years ago this is easily understood The author of the novel containing three love triangles suffered a similar problem The author did not like the double prime for the opening and closing quotes When he wrote the first part he selected special graphics characters for the quotes When he wrote the other two parts the smart quotes were automatic for him He did not recall why it may have been a new revision of his formatter 4 3 Not using available functionality In two of the test ca
14. r Writer2LaTexX March 2012 http sourceforge net projects writer2latex Bart Childs Texas A amp M University College Station TeXas 77843 3112 USA bart at tamu dot edu http faculty cse tamu edu bart BTEX source from word processors
15. s the intent in creating Writer2LaTeX to give the user ATEX as an improved output device I think that poses a bigger challenge How do you teach a Writer user to write for ATRX 9 Questions I did not intend this as a FAQ but thought it might be a good way to end the present paper LL FTX Do any of the test cases use ATEX be yond Leslie Lamport s book Answer No for memoir and book on the three love triangles Yes for the science and engineering texts Packages used float lscape makeidx fancyvrb graphicx array amsmath amssymb sidecap wrapfig and caption These were probably not all necessary but useful 183 Word test case What do you want for a Word test Answer A one pager like Norman Naugle s An El ementary Sum Then many others would help I hope it would also convert to Writer and back too How long How long did it take you to type Nor man s note Answer An hour or so The next question might be why didn t you just do it in Word Well probably that would have taken seven or eight hours and fortunately I do not have Word in my house References 1 Charles Ota Heller Prague My Long Journey Home Abbott Press December 2011 2 Wilfried Hennings Converters from PC Textprocessors to ATRX Overview June 2012 http tug org utilities texconv pctotex html 3 Karel Hor k Those obscure accents TUGboat 29 1 42 44 2007 4 Henrik Just User s manual fo
16. ses the authors used itemized lists The exported form yielded consecutive lists of one item This did not bother the bulleted lists but would have been an error with enumerated and description lists In many cases the authors did not use styles so chapter and section beginnings show the formatting but no ATEX commands This is not a total loss because I convert the section numbers into labels that would aid if we were trying to resolve differences in my output with the older version 4 4 Oops These examples can be difficult A glaring example is that WordPerfect s mathematics operators may follow the operand in some cases In ATEX the Bart Childs TUGboat Volume 33 2012 No 2 operator is always first I did not find a general rule as to when to expect this My Emacs macros for adjusting this are interactive to enable the user me to minimize such problems A really big oops worth repeating is the lack of using styles which caused inconsistencies I had to handle some of these manually 5 Typical Emacs macros The first versions of these macros were developed when I was using an export that was usually desig nated tert with encoding This export would discard all or nearly all formatting such as emphases The improvements in Writer2LaTeX have led to a re duced need of this kind of detailed editing Still the concepts in the design of these macros are applica ble in the current system of conversion as well as key
17. sive white space and formatting trash This almost always includes many of the grave characters this must be an escape character for the internal form of Word Perfect mathematics I have not had a reasonable test case with Word mathematics yet There are small examples of math ematics in the programming text 5 3 Display mathematics The concepts in the previous subsection are appli cable However there are several forms of display mathematics These forms were used in the rotordy namics text 1 J the standard for display equations without numbers 2 beginfequation end equation just an alias for the former or vice versa 3 beginf equation end equation which numbers the equations and should have an ac companying label 4 begin equation begin split end split end equation which numbers a collection of equations and should have an accompanying label 181 Chapter 8 of Frank Mittelbach et al s ATEX Com panion has some seventy pages of excellent details of advanced mathematics formatting I implemented these four display math choices using one function key and prompting the user for which of the above forms was desired I developed similar choice macros for presenting fractions and matrices which made conversions faster and most importantly more consistent The most important facet of this conversion is that with a little care the totality of the mathematics was converted correctl
18. two 1 Genesis My primary formatting system has been TRX based for more than thirty years Throughout this time I have had occasional need to import small parts of documents done in word processors into my TEX based documents I have accomplished that in a number of ways from keyboarding small projects to somewhat automatic conversion depending upon what was available I used some of the earlier systems discussed by Hennings 2 Several years ago two colleagues were writing a text on programming and became aware that they would have significant advantages if they could convert the half of the book that was completed to BTFEX and take some instruction on how to complete the rest in ATEX I sketched the process and created a small set of Emacs Elisp macros to do that conversion We agreed to the generalities with plans to make a formal agreement upon the return of the senior author from Bart Childs TUGboat Volume 33 2012 No 2 a summer long trip Much of the ATEX work was to be done by the junior author naturally The health of the junior author suddenly deteriorated and my conversion project was cancelled I continued to be intrigued by the concept I learned more Elisp added macros and a number of open packages that seemed to offer promise as a means of getting much of the conversion done in an automatic manner I never felt that a mostly auto matic conversion was realistic for projects involving significant math
19. y and hours of detailed laborious proofreading was avoided 5 4 Programs code fragments verbatim text Programs should be formatted by language sensi tive packages like listings The package fancyvrb requires some study but gives great results Both packages come with inline commands whose use is aided by adaptation of the above font changing and inline mathematics concepts 5 5 Other macros fix up There were several other macros that aided the con version I consider these to be fix up in nature These include e captions in the rotordynamics text often con tain inline mathematics The use of the ATEX delimiters C is not allowed they must be converted to the TFX toggle e Interactive aid to standardizing presentation of fixed point and floating point numbers e Locating multicharacter super subscripts that were likely exported incorrectly needing group ing e Locating likely problems due to insertion of in advertent whitespace e Locating unescaped TEX control characters e Macros to aid the insertion of labels and their references 6 Current system The current system has been improved greatly with the release of OOo Writer2LaTeX version beta 1 2 Despite its being labeled a beta release I have not found any problems to date I find these observations about this new release interesting 1 the user s guide is 10 shorter and 2 the output files are 3 5 shorter than with ver sion 1 0
Download Pdf Manuals
Related Search