I thought that a blog post would be more fitting for my response, so here we go. Below, feedback from Brian is in blockquotes (gray background), and my responses are below that.
Automatically disable Glossary searches during Auto Translate
(unnecessarily slows the AT process down).
This is a good point, especially if you’re using fuzzy glossary matching, which can be slow. The reason why Felix does glossary lookups during auto translate is because of an implementation detail that doesn’t really matter; it will be fairly easy to change it, so I’ll plan on doing so within the next few minor releases.
Ability to run through cells vertically in Excel (instead of just horizontally).
I’ve got a ticket for this, with a planned fix of releases 1.6.5 (the upcoming release is 1.6.4, due within a week).
Ability to skip hidden cells (rows/columns) in Excel. Not just for Auto Translate, but also when going through cells one at a time.
As you noted, I recently changed the Auto Translate functions so that they skip hidden cells. At the time, I didn’t make it a rule to skip hidden cells in general, because I wasn’t sure if I should make that a user option. I’ve convinced myself that it doesn’t need to be a user option (you could just tell the user to unhide the cells if they want them translated), so I’m now planning to skip hidden cells in all cases.
Ability to specify characters to ignore (assuming entire character string is made of up ignore characters). There was a separate post for this a few months ago.
This is another feature that I have a ticket for. To step back and talk about the big picture for a moment, I’m planning on changing the preference interfaces for MS Office/Excel/PPT. The first look at what the new interfaces will be like can be seen in the new shortcut customization/abbreviations list dialogs.
Therefore, what’s been holding up this feature has been the implementation of the new interface. Once the new preferences interface is done, there will be a cascade of new customization features for MS Office. I’m planning on doing this change by release 1.6.5 or (probably more realistically) 1.6.6.
Perhaps this is a MS limitation, but using Felix in a textbox in Excel, allow for running through individual sentences (segments) instead of only the entire textbox. Larger textboxes with lots of sentences get chopped off.
Yes, this is due to a rather arcane “feature” of MS Excel (follow the link if you’re really interested in such things). There is a workaround, which I hope to implement within the next few minor releases.
In Excel, ability for Felix to include textboxes when running through document (both manually and Auto Translate). Separately clicking individual textboxes one-at-a-time can be cumbersome, particularly for files almost entirely comprised of textboxes (can sometimes literally be hundreds of textboxes).
I’ve got a ticket for this in both Excel and Word. There are a couple of complexities with this, which is why I’ve been pushing it back for a while now, but I hope to get it implemented soon.
A way to have Context entered globally as you work in Felix.
This is another ticket, and I’m planning to have it done by the next release (1.6.4). One of the reasons for the new Memory Manager interface in Felix was to hang on new options like this. Soon, you’ll be able to go to a TM properties page and set a default context.
Store only text data in the TM, not additional font formatting data, and have font formatting data used for that particular segment only. I’m noticing that I’m getting font formatting from previous jobs inadvertently inserted into subsequent jobs that I don’t want the formatting carried over.
You can do this now. In Word or PowerPoint, go to the Felix menu, and then Felix Properties >> Format, and clear all the format-type checkboxes.
The questions below relate to Memory Serves, which is a free program for sharing Felix TMs and glossaries via a server.
In Memory Serves, allow Replace feature to distinguish between either Source, Translation, or both. Does not allow searching and replacing of specific field (i.e. I may want to replace text only in target language).
You can do this now using tags. (source: for source and trans: for translation).
Example: Replace “apple” in the source with “banana”.
From the replace page, click “Replace Help” for a list of all the tags, and some examples.
Expanding search/replace items further, I want to be able to search by specific field (e.g. all entries modified by a specific user). For example, Ted is a new user. I’ve noticed that almost all of Ted’s entries are incorrect. I want to be able to see all of Ted’s entries and I also want to be able to delete all of Ted’s entries as needed.
Right. As above, you can use tags to specify various things, including who created the entry, when it was created, and so on. From the replace page, click “Replace Help” for a list of all the tags, and some examples.
I’m also planning to add a feature called “delete all matches,” by the next release of Memory Serves, and within one or two minor releases of Felix. This will be along the lines of the also recently added “save matches to TM/glossary”.
As a former user of Wordfast, I’ve got a lot of segments with extraneous code. For example, “&’B;” (Wordfast coding for line break), “&’BB;” (Wordfast coding for katakana “sa”), among others. Is there a way to search/replace the line break string with a line break in Felix?
You can do this in Memory Serves now, using the “regex” tag.
Here’s an example:
Felix also supports the regex tag, but in this case you would need to use the HTML tag “<br />”, which will cause a newline to be added. I’ve filed a ticket to support the “\n” tag as well.
[Bug?] When edits are made to records in Memory Serves, “Modified by” field shows computer name of PC running as server, not the user name.
Yes, this is a bug. It should be set to the name of the logged-on user, and fall back to the computer name if no one is logged in. From Felix, it should use the Felix user name.
Does memory search speed (when performing Look Up command) slow as the number of entries increases? What about when multiple memories are loaded and active in Memory Manager? Is it faster/slower to search
100,000 entries in one memory or 100,000 in two or more memories?
Yes, the search slows down with the total number of entries, but not the number of memories (unless you’ve got something ridiculous like thousands of memories). Therefore, speed shouldn’t be a consideration for deciding whether to split up TMs.
I’ll also note that since you can save and load preference files, with all the currently open TMs and glossaries, you can use multiple TMs and glossaries, and then load them all just by dragging and dropping in the prefs file (or going to Tools >> Load Preferences).
Two things that tend to slow down searches the most are very long query sentences, and large glossaries with fuzzy gloss matching enabled.
I’d like to thank Brian for the great feedback, and encourage any user with feedback, suggestions, or problems to let me know. In addition to the mailing list, there is a forum (Japanese forum here), and you can use the contact page and email me directly, or leave comments in the blog.