alexpgp: (Visa)
I got to wondering the other day about how to go about delivering a job for which, to make things easier, I've combined a number of related files into one Word "master document."

I still haven't figured out the answer, other than to find advice along the following lines scattered throughout the Web:
The Master Document feature in Word remains word processing's answer to Conan the Destroyer
There are two kinds of Master Documents: Those that are corrupt and those that will be corrupt soon.
Lucky me.

Oh, well... I'm off to split the file! Scramble. Scramble. Scramble!


Where did I get the idea I had to "split" anything? I came to this conclusion after opening one of the files that was part of the master document, only to find that the translations and edits done to the document were not there.

In my haste to make sure my work would not be lost in the event of a "master document corruption event," it did not occur to me to wonder where that work was stored (as it had to have been, since everything came up properly upon opening the master document).

This morning, it did occur to me, and the answer was quite simple and a result of inattention on my part.

It turns out the original files that are part of the master are "doc" files. Apparently, when the master is saved, all of the files are saved as "docx" files, which is where all the changes are. (I knew there had to be a logical explanation somewhere!)

What I had done earlier was open a "doc" file instead of a "docx" file. The only mitigating circumstance here (assuming you'd call it that) is the fact that the job contains a bunch of files with very similar names, so I just clicked on the first file having an appropriate name, which happened to have a "doc" extension.

My current solution is to make regular incremental backups of the directory in which the constituent files are kept, and it will be these constituent files that will be sent to the client.


Jan. 15th, 2013 10:43 pm
alexpgp: (Default)
I haven't been listening too closely to any telecasts, but Galina told me the weather forecasters are calling this the coldest January 15 the Houston area has experienced since 1917. I believe it. Walking the dogs was no fun, for either me or the dogs.

The Routine Stuff™ was fairly light this morning, so I spent the bulk of the day despeckling, getting about 22,000 words into their go-to-meeting finery and out the door. There's about 36,000 words left in the job, and I hope I can get a good chunk of that done tomorrow (along with a translation of a 1,500-word author's bio that is, if anything, more complex than the rest of the material). I'm looking forward to invoicing the job, that's for sure. I represents a bunch of time, spent here and there when things were slow over the past couple of months.

The paperwork for the new LLC arrived today. Besides whatever needs to get done tomorrow, I'll need to review it and then make sure all the details are taken care of, so I can formally announce a name change to my clients.

* * *
My Word-fu got a boost today during the despeckling process when I had to figure out a way to look for words like образ and орбразец (and their declined forms, e.g., образа, образу, образом, образе, образы, образов, образам, образами, and образах just for образ) while avoiding having to slog through variations of the word образование during the search.

I ended up creating a search string that used wild cards:
which basically specifies a search for "обра" immediately followed by between one and five lower-case Cyrillic characters, followed by 'end-of-word', which is denoted by the '>' sign. Without that '>' sign—and I'll admit I learned this the hard way—the search string would still stop at the frequently encountered образование (hey, there's "обра", immediately followed by five Cyrillic characters, immediately followed by even more Cyrillic characters, y'dig?).

Live and learn.



Jun. 23rd, 2010 07:42 am
alexpgp: (Computing)
Problem: You have a one-column table having a fixed and large number of rows, where some of the rows contain information to be translated. The row above each row with information to be translated has been left blank. The final product shall display the translations in the row immediately above the row in which individual pieces of information are located (i.e., the final product shall be bilingual).

Using a translation memory tool such as Wordfast, one can easily arrive at a result consisting of a translated table, i.e., given the following source table:
source A
source B
source C
it is easy enough to obtain this:
target A
target B
target C
The question is, how to get this:
target A
source A
target B
source B
target C
source C
Solution: Remove the blank row at the top of and add a blank row at the bottom of the target table. Copy the result and paste it as a column next to the source column. This is the result:
 target A
source A 
 target B
source B 
 target C
source C 
Then convert the table to text, making sure there are no paragraph marks in the table (if there are any, search and replace with some innocuous character, e.g., '^') and separating text with a character, e.g., '@', to get:
@target A
source A@
@target B
source B@
@target C
source C@
Now convert this text to a single-column table, separating text at paragraph marks, and then use search-and-replace to get rid of the '@' characters. The result will be the desired table.

Half of this job isn't knowing the language, it's using the tools.



alexpgp: (Default)

October 2017

1 2 3 45 67
8 9 10 11121314
15 161718192021


RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Oct. 17th, 2017 04:53 pm
Powered by Dreamwidth Studios