Denis,<br><br>I really appreciate your comments. I'm hoping that with improvement it can become a high-quality choice available for the interface font.<br><br><div><span class="gmail_quote">On 5/5/07, <b class="gmail_sendername">
Denis Jacquerye</b> <<a href="mailto:moyogo@gmail.com">moyogo@gmail.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
- At first I got the impression g stands out a bit much, it might be<br>too original with such an open lower bowl, although I eventually got<br>used to it.</blockquote><div><br>The g is a bit funky. There's an alternate g (looks like a q with a tail) in each of the font files. You can view it with FontForge. I'm not sure what the best way of handling alternates is. Maybe have a FontForge script that asks people which glyph they like better (I have a series of alternates as you'll see later), and then automatically re-encoding the font?
<br><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">- I find the bar of the f a bit short on the right in Regular, it's<br>almost the same length as the bar of t in Bold which looks more
<br>balanced.</blockquote><div><br>I'll have to look at this. For a long time, I had the bar on the t too short, until someone on Typophile pointed out just how short it was. In part this is a reaction to Bitstream Vera/DejaVu, which has ridiculously long bars.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">- There's something odd with v, w and y in Italic. Maybe v, w should<br>be more similar to y, or vice versa.
</blockquote><div><br>Definitely weird. :) These are experimental glyphs. I really like the v, but the w and y both need work. Look in the file for an alternate y. It fits better with the v and w, but it feels like it's falling over too much. The current y is inspired by Palatino's italic y. With the more vertical sides, it seemed that the flag on the top right closed the counter too much, so I removed it.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">- Diacritics are a bit much to the right on C and G. But looking at<br>U+0134, U+013D and U+01E9 with off diacritics, composed glyphs
<br>probably need polishing too. I suggest placing anchors on a visually<br>centered axis and then build accented accent through Fontforge. This<br>way it will be consistent through pre-composed glyphs and with<br>composed characters.
</blockquote><div><br>Yeah, the diacritics need a lot of work. Essentially they were just thrown up there so that there wouldn't be these funny holes in multi-lingual webpages. The difficulty with using FontForge is that I'm developing the font in MetaType1. It's great because I can compute (based on existing points, etc) the location for the visual axis. MetaType1 can then compose the glyphs. Unfortunately, MetaType1 outputs a Type 1 Postscript font which I then import into FontForge, and I don't think that there's any way to store glyph placement info in a Type 1 font (maybe I'm wrong?). Is there an advantage to having FontForge build accented glyphs, as opposed to having MetaType1 do it?
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">- Combining diacritcs should probably be centered on X=0 and have zero<br>advance width, at least if you want to follow the OpenType definition.
<br>Otherwise, they should have their contour in the negative range for<br>legacy system that cannot use anchors to position them. This doesn't<br>apply to the Mono font, where all characters need to have the same<br>
advance width, but you can use OT features to fix it.</blockquote><div><br>Same difficulties as above.<br> <br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
- When you do polishing, remember to removed duplicate points, there<br>are a few like the top point of 'a', </blockquote><div><br>This seems to be a problem with FontForge's simplify command. I'll write to George about it.
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">also remember to round coordinates and bearings.</blockquote><div><br>How important is this? I've found that rounding distorts the shapes. However, freetype seems quite happy with the fractional coordinates, and it displays fine. I haven't noticed any other problems with it (displays fine in pdf viewers), but I haven't used the font much besides on-screen.
</div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">If you want to make it easier to interpolate weight variant fonts,<br>make sure each glyph has the same number of point and the same order
<br>in Regular and Bold.</blockquote><div><br>Since I'm using MetaType1, it can do all of the "meta" stuff itself---that's one of the advantages of using it.<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Let us know if/how we can contribute glyphs or features.<br></blockquote></div><br>Thanks! Comments like these are quite helpful. I'm in the process of figuring out the best way to release the MetaType1 source. It seems that it would be better if changes are made to that rather than graphically changing the glyphs in FontForge. It also raises the interesting question of what license is most appropriate, since there is source code.
<br><br>Best wishes,<br>Stephen<br><br>