pessimism

I just think the chosen background colour is a little too dark. Something about the colour (#F5F5F5?) is off to me.

The new design is really great and in many ways better than GitHub’s. Now people unfamiliar with the services might actually be able to find a download button for once.

Another place where GitHub is really dropping the ball is in the social aspect of their service with terrible activity feeds I can’t believe anyone actually uses. I wrote a rant about it here: http://pygm.us/uGhNdcGU.

There are many, many ways Bitbucket can beat—and beats—GitHub, so this new design makes me a lot more optimistic about the continuing competition between the two and the improvements this will result in on both sides.

I will say that I’m not sure whether I like a grey as dark as the one you use in your new design, though. :)

On that note, does anyone know a working S3 policy script to prevent hotlinking outside your domain/bucket? I can't get any to work.

Well, ask if people are satisfied with their own blog CMS, and they'll probably say "no". I had to use something as obscure as Blogofile to satisfy my needs.

## Technically

I think some competition for Jekyll is sorely missed, and no one who isn’t a nerd even knows about static blogging. Granted, it's also been a royal pain for me, when it comes to pushing automated tweets, whenever I update my blog, so there are a lot of problems to be solved. I still haven’t bothered to write something in Python that tweets based on my RSS activity. (ifttt hasn’t worked optimally for me.)

Nevertheless, it’s liberating to be able to simply create a Markdown file and write a post like this:

    ---
    title: Something Something Paul Graham
    date:  2012/09/30 23:00:00
    ---
    That Paul Graham might have a point.

Run Blogofile’s Amazon S3 uploader with one command, and boom, uploaded. As a bonus, I can back up my blog really easily using version/change control, since all files are static.

There are a lot of steps here too complicated for the average user, but all it takes is a moderately simple Mac app, and most people are golden - especially now that everyone has Dropbox to store their files.

It seems that there is a tendency for people who want to revolutionize blogging and social media to reinvent RSS. I doubt that’s the way forward.

---

## Socially

If I were to suggest an idea for a Wordpress-like platform, it would be a reinvention of the blogroll. When people don’t have the reach to attract advertizers, they can instead link to other users on the same platform and vice versa for exposure.

I would find a way to gauge the reach of blogs, and allow customers to find bloggers with similar reach to collaborate with for a mutually beneficial exposure that attracts traffic. I imagine a business model could be built on this alone, but it would probably be much easier, if people were on the same platform, say, Tumblr or Wordpress.

Tumblr really revolutionized blogging by adding a community layer onto their platform to make it a social experience that married Wordpress blogging and Twitter.

The profile is a real improvement (although it might be preferable to fit more repos into a smaller space?), but the public activity stream still seems as hopeless as before. I pointed out some reasons with pictures for the older version, but the points still apply: http://pygm.us/RFLFk4nt.

Thanks for the typeface. I really appreciate it.

If you want an example of it in action, you can go to a deployed example of my app at http://pygm.us/kzXSwSRV with the log-in admin//password. I haven't gotten around to letting people change their settings, but you can use the button in the sidebar to see it in effect. If you apply the CSS in one of the threads, you may see why I find the font weight to be a bit on the heavy side. But to each their own and all.

Of course, now that Instapaper supports the typeface, you should check it out in it as well, if you have an iOS device.

Thanks again! :)

As someone working on a free open-source project that is almost as text-intensive as it can get, I was really happy to see a dyslexia font/typeface that was truly free, when Marco Arment announced it would be included in Instapaper.

The gist of the research is to create typefaces whose characters, even flipped and rotated, cannot be confused at all.

I've seen others like http://redd.it/j277w that, as interesting as the look, weren't free, and as a result come with a lot of problems.

There are two approaches:

1. Client-side typeface support.

2. Service-side typeface support.

People with dyslexia can easily purchase a dyslexia typeface and embed it in their OS and force their browsers to display websites with the new typeface, while service-providers can't always just fork over five bucks and link to the .ttf or .otf file without getting into trouble.

Paid typefaces are great for clients, but aren't always for service-providers.

I have always wondered why Apple have never provided a dyslexia typeface with iOS and OS X. Considering their efforts in accessibility, maybe they don't believe in the efficacy of dyslexia-friendly typefaces.

While there is still a discussion of quality[1] and efficacy to be had, the real news here is that someone made the effort of providing people with a truly free dyslexia typeface, which has been sorely missed to this day.

[1]: I think what irks me is the large font weight even for the regular font. The bold font is barely discernable from the regular.

Here is a personal experience on what it can be like as a customer of a product available both in the MAS and outside it.

A while ago, I tried the Fantastical trial, and when I used the purchase link, I was immediately prompted to a credit card payment screen. I paid and bought the software, and all was fine and well.

But later on, I learn that the software was also available in the MAS - something Fantastical didn't really bother explaining to me. When I contacted them for any kind of refund (to be used to purchase the app in MAS), I was told that they couldn't accommodate that since a lot of other users have requested the same thing - which already seems like a weird way of thinking.

They price the software in both places the same - meaning they earn more money, if people don't by it in the MAS, and I felt the entire experience to be very deceptive.

All in all, a really bad experience that I can't recommend any developer to imitate, as it does nothing to inspire customer loyalty. Especially when the developer has a financial incentive to not addressing the issue. At least from my limited perception as a non-Mac developer.

Again, this is the impression you get as a customer. I'm not saying this is the intent, but perception is reality in the real of customer experience.

Consider this another con to double-dipping, if you intend to do something similar.

Their customer is abysmal. They never respond on Twitter—fair enough, other registrars do, though—and it took four days to get a response to why I couldn’t log in. I even had to go through weird escalations to higher “tiers” of support.

If someone knows a good non-American registrar with a wide selection of domains who also has two-factor authentication like Name.com, please let me know. Until then, I’ll probably just postpone the domain purchases.

Speaking of gandi, a while ago, I ran into some problems logging in to their site, and I—finally—got this response from them by e-mail:

    Hello again pessimism,
    
    Our password field supports only passwords up to 16
    characters at this time. All longer passwords
    are truncated.
    
    If you have any further questions, please let me know.
    
    Sincerely,
    
    [redacted]
    Tier 1 Tech Support
    Gandi US

I checked in with @theharmonyguy who says this is a pretty bad thing, and I originally intended to do an Ask/Tell HN post about it, but life got in the way.

What are your thoughts on this and its implications for using gandi?

My mind was slightly blown today, when I realized that http://productivitycalendar.com and .net (and all other permutations) were available.

There are still some excellent domains left, as difficult as it can seem to find some good ones.

Let this be some anecdotal pep talk.

Using NLTK to "mine" for viable domains also sounds like an interesting idea. Might even be a business model.