#Texpad real laggy pdf#
#Texpad real laggy code#

Undoubtably a change this severe will be painful for some, but it will be less less painful than heading out to the computer shop in 3 years time to purchase the 2TB harddrive that will be required for the exponentially expanding tex updates. Once again, the time freed up by culling packages would more than compensate for the time required to fix up the knockon effects throughout the TeX system. pstricks and tikz, or bibtex and biber) make a choice and keep one. With just one underlying typesetter to focus on, development of that typesetter would accelerate greatly and many packages would become obsolete, further freeing up time for purposeful TeX development.Ī second phase to this reorganisation would be to carry out the same trick with the packages.

#Texpad real laggy pdf#

XeTeX outputs directly to pdf in one step, it is by far the easiest tool to support for external products, its usage of unicode and integration with modern fonts would allow the ancient texmf to be dropped entirely, and with a little bit of development platex (japanese character support), could become unnecessary. Those are my complaints, so what is my solution? Texlive 2013 (2012 is frozen at the end of this month) should have just one typesetting tool - XeTeX. More importantly if the community’s endeavours hadn’t been split between all five typesetting chains, they could have been spent upgrading one typesetting chain to handle all contingencies and making that one tool truly great. The effort we will put into supporting Biber could go into a tikz GUI now that pstricks has been replaced. For example, the effort we have put into supporting the four surplus toolchains could equally have gone into better autocomplete, or a pstricks GUI. The true problem with this fragmentation is that the effort put into LaTeX and its ecosystem has been split. This is more a symptom than it is a concern in itself, but an increase 1.6GB in four years isn’t moderate bitcreep, it is a sign of morbidity and it is unsustainable. TexLive 2007 takes up 1.37 GB on my hard drive, TexLive 2008 requires 1.76GB of space, but by TexLive 2011 the installation has grown to a stunning 3.01GB. For all its advantages, Biber has further fragmented the LaTeX project.Īnother effect of this fragmentation is bloat, which is very straightforward to measure. Biber, the replacement for bibtex in the biblatex chain, is a great improvement over the venerable bibtex, but now there are two biblatex options, both actively developed, but only one of which is necessary.

#Texpad real laggy code#

This fragmentation of code and development effort has recently spread beyond the core typesetting tools. Unforunately, so are the efforts of everyone else putting time and money into LaTeX. Instead of supporting one typesetting chain, we have had to support five and our efforts have been split. On top of that there is XeLaTeX for unicode fonts, and pLaTeX for an alternative approach to multibyte characters often used in Japan. However, most people don’t want an unnecessarily slow 3 step process, so we support the single stage pdfLaTeX as well. This package will only work when there is a PostScript intermediate in the typesetting chain, for example the LaTeX -> dvips -> ps2pdf chain. That explains the need for pdfLaTeX and LaTeX->dviPDF, but why do we also support LaTeX -> dvips -> ps2pdf? One reason is the pstricks package. In this typical example of LaTeX’s fragmentation there are two typesetters and two associated packages in the place of one single working typesetter. Not to worry, there are add on packages to convert JPEG or PNG to eps so that LaTeX can use them, and likewise there is an EPS to PDF conversion tool for pdfLaTeX. LaTeX on the other hand includes eps but not JPEGs or PNGs.

pdfLaTeX can directly include standard image formats such as JPEG and PNG, but it needs an eps2pdf converter to directly embed EPS content. The obvious question is why? why do we need five toolchains performing exactly the same task? This is easiest explained with the problem of embedding image files. Maintaining and updating Texpad over the past six months has given us insight into the workings of the TeX system, the problems it faces, and its potential saviour - XeLaTeX.Ĭurrently we support five typesetting toolchains in Texpad, pdfLaTeX, LaTeX -> dviPDF, XeLaTeX, LaTeX -> dvips -> ps2pdf, pLaTeX->dvipdfmx. We took we learned during this project and built a desktop LaTeX editor, Texpad. A while back I wrote about the unfortunate state of the LaTeX codebase, and how it scuppered our project to port it to the iPad.