To make a short (two page) résumé in PDF format, use make. To see the full CV including social media and committee-ships, use make longform.

Trying something different here, let's make two files: a long-form CV with all the details and a short-form résumé that just puts the exciting stuff on a single page. Both are generated from a single LaTeX source file.

It's not straightforward to pass parameters from the command line to TeX to accomplish something like conditional compilation in C. Here we use the most reliable technique, where the LaTeX source file unconditionally \inputs a file that is dynamically created by the Makefile just before it's needed.

The résumé is a single page and not very full at that. It shows the bare minimum information a harried screener wants to see:

• What positions, at what companies, has this person held?

• Did he or she do anything interesting there?

• Phone number and email address for the candidate.

To be nice to people trying to process files with automated tools that might not properly handle non-ASCII characters in filenames, the file is called resume, not résumé.

This file is about four pages long in PDF. The first page is the résumé. Keywords are relegated to the last page where automated résumé filters will still find them, but they won't distract human readers. Keywords are separated by commas and formatting is deliberately kept simple:

Certification and Accreditation (C&A) of cross domain systems, Assessment and
Authorization (A&A) for classified environments, Certification Test and Evaluation
(CT&E), Security Test and Evaluation (ST&E), penetration testing, DIACAP, Common
Criteria, DCID 6/3, programming in C and assembly language with a strong
interest in Lisp derived languages, Compartmented Mode Workstation (CMW) programming,
UNIX, Trusted Solaris versions 2.5-8, software development, R&D, principal
investigator, U.S. citizen with TS/SCI clearance, technical writing, CISSP-ISSEP
for DoD 8570.01 IAT, IAM, and IASAE Level III compliance required for software
development that crosses multiple enclaves, experience living overseas, programming,
publications, patents, public speaking, funding, research, and teaching experience.

Keywords are in no particular order, just arranged for even line length and colour when printed. I thought about making this section invisible, the way SEO companies put meta keywords in web pages to get them indexed by search engines (and spammers use obfuscated text to evade mail filters) but decided it would be dishonest, like malware in a PDF file.

Ligatures like fi and ffl look great on the printed page but they can be a barrier to copy & paste. Some PDF readers, notably Adobe Acrobat X, don't handle ligatures correctly when pasting to plain text. The word Certification comes out Certi cation and that doesn't match a naïve regex. The solution is to load the cmap package in LaTeX like this:

\usepackage[resetfonts]{cmap}

The resetfonts option is required when used with the Computer Modern fonts which by default are a little too clever with ligatures. Remember, the goal here is to make things as easy as possible for keyword filters used by Human Resources. The same CMAP tables can do other useful tricks like transforming curly quotation marks to straight ones upon copy to plain text; “proper quotation marks” tend to paste into plain text as Unicode, and that can also confuse a straightforward keyword search. (It depends on the PDF reader; Adobe Acrobat Reader X doesn't understand ligatures at all and pastes curly quotation marks as unicode characters, but searches correctly across hyphens; in Mac OS X, the Preview programme gets ligatures right and its search function correctly matches quotation marks whether curly or straight.)

To make sure grep finds all my keywords, I altered the default CMAP table that gets compiled into the PDF output, replacing every instance of <201C> (Unicode left double quotation mark) and <201D> (Unicode right double quotation mark) with <0022> (ASCII double quote) and all instances of <2018> (Unicode left single quotation mark) and <2019> (Unicode right single quotation mark) with <0027> (ASCII single quotation mark). This makes single and double quotation marks in the PDF file copy and paste as straight quotes. I further caused en dashes to be replaced by a single hyphen and em dashes by two hyphens, to match traditional typewriter usage. The ot1.cmap file is written in PostScript; to override it, I simply made a copy in the current directory and modified the copy:

$ diff ot1.cmap.original ot1.cmap
21c21
< 8 beginbfrange
---
> 7 beginbfrange
29d28
< <7B> <7C> <2013>
31c30
< 40 beginbfchar
---
> 42 beginbfchar
61,62c60,61
< <22> <201D>
< <27> <2019>
---
> <22> <0022>
> <27> <0027>
66c65
< <5C> <201C>
---
> <5C> <0022>
68c67,69
< <60> <2018>
---
> <60> <0027>
> <7B> <002D>
> <7C> <002D002D>

This avoids having to install any software in the filesystem; PDFLaTeX preferentially uses the modified ot1.cmap file if it finds one in the current directory.

A similar trick is done with LaTeX's bulleted lists, which use the math mode \bullet because bullets don't appear in TeX's OT1 font encoding. The following modifications to the oms.cmap file in the current directory cause bullets at all four levels of a LaTeX itemised list in the PDF file to automatically turn into asterisks, hyphens, or plus signs when copied and pasted to plain text.

$ diff oms.cmap.original oms.cmap
28c28
< <01> <00B7>
---
> <01> <002B>
42c42
< <0F> <2219>
---
> <0F> <002A>

I had to modify the distribution version of cmap.sty to get it to work with the 11pt and 12pt options to \documentclass as follows:

$ diff cmap.sty.original cmap.sty
18,21c18,21
<     OT1/cmr/m/n/5,OT1/cmr/m/n/7,OT1/cmr/m/n/10,%
<     OML/cmm/m/it/5,OML/cmm/m/it/7,OML/cmm/m/it/10,%
<     OMS/cmsy/m/n/5,OMS/cmsy/m/n/7,OMS/cmsy/m/n/10,%
<     OMX/cmex/m/n/10%
---
>     OT1/cmr/m/n/5,OT1/cmr/m/n/7,OT1/cmr/m/n/10,OT1/cmr/m/n/11,OT1/cmr/m/n/12,%
>     OML/cmm/m/it/5,OML/cmm/m/it/7,OML/cmm/m/it/10,OML/cmm/m/it/11,OML/cmm/m/it/12,%
>     OMS/cmsy/m/n/5,OMS/cmsy/m/n/7,OMS/cmsy/m/n/10,OMS/cmsy/m/n/11,OMS/cmsy/m/n/12,%
>     OMX/cmex/m/n/10,OMX/cmex/m/n/11,OMX/cmex/m/n/12%

(The TeX font encodings are documented here.) The resulting plain ASCII when pasted is 7-bit clean and displays correctly no matter what international character set encoding is used on the recipient's machine.

For some reason this works at with the 10pt and 12pt options in LaTeX, but not 11pt. Diffing the log files doesn't suggest any obvious reason why that should happen. For now, I'm just going to rewrite the résumé to look good in 12pt type. It's easier to read for the hiring manager in that size anyway.

Embedded fonts in PDF files can cause problems later if displayed on a system that doesn't have the expected fonts installed. To me sure of embedded fonts in PDF files, do this after building the file:

ps_file = $(target).ps
#
# Now, make sure embedded fonts won't cause a problem for somebody else.
#
pdf2ps $(pdf_file)
rm -f $(pdf_file)
ps2pdf -dPDFSETTINGS=/prepress $(ps_file)
chmod a-x,a+r $(pdf_file)

It makes the PDF file bigger, and seems to make TeX fonts lighter in weight visually, and hyperlinks in the PDF file disappear, but the resulting PDF file will have all fonts embedded. It breaks the metadata that I like to insert in PDF files: author, title, key words, et cetera. Interestingly, the names of fonts are lost; when viewed in 'Properties' in Adobe Acrobat Reader, the fonts in the PDF file have names like 'T3Font_154' (Type 3) instead of 'CMBX12' (Type 1).

It seems impossible to turn off all the PDF security options; 'document assembly' is always disallowed, although I wonder how many PDF display and manipulate programmes actually follow it, like Mac OS X preview for example, because I've never had problems inserting pages into PDF files because of it.

Not recommended to use it for those reasons. There is no point putting visible hyperlinks in the PDF file only to disable them.

PDF is basically PostScript under the covers and we can do some things to make the resulting binary file easier to parse. Side-by-side columns can be problematic for parsers; they tend to interleave their text in a copy and paste to plain text, or a grep. Small caps render correctly from PDF into plain text, as mixed case, I have found.

Make targets include the default all to emit the short-form CV (two pages), longform to emit the three-page version of the CV including social media logins and committee-ships, edit to quickly edit the source file, resume to emit the short form résumé, clean to remove temporary files (before commit), commit to update the local Git repository, sync to push changes to GitHub and merge any non-local changes, allclean to remove everything that can be regenerated, help give a quick overview of options, and spell to check spelling against a UK English dictionary.

% make

For information contact the author: [email protected]