One thing about abbreviations in Arabic is they will be read as a word, not as distinct letters. In English, "http" can't be read as a word, you always read it as "ech-tee-tee-pee". "etc" is not read "etk" but read as "ee-tee-see". Where as in arabic, "الخ" is read as one word "ilkh" instead of as letters "alif-lam-kha".
So, to abbreviate words, you have to think hard about how people will read it. For instance, someone mentioned in another reply "Hamas" and "Fatah". Both these abbreviations read nicely as other words with related meanings. But, how would one abbreviate "Filisteen"?
If you choose the first three letters "Filis", well, that reads like "penny" and is generally used to refer to cheap things - you don't want to use that.
If you use "ft", that would read as "FiTT" فط which just sounds awful in Arabic. (it's not like "fit", the "t" here is heavy/strong).
Actually, "filisteen" is already bad enough: it sounds like "filis" + "TTeen" which is "penny" + "dirt" (we used to make fun of this when we were kids). Though in Palestinian (and the Levant region in general) they pronounce it "falasteen", not "filisteen".
It's pronounced "et cetera" because that's what it's an abbreviation of. Most abbreviations (unlike acronyms) aren't verbalized in their abbreviated form. You say "versus", not "vs" when you read it.
However, I do often hear Unix hackers say "et-sey" when referring to the "etc/" directory because precision is important there: the directory is not named "etcetera" so pronouncing it as such would confuse.
It must depend on your definition of "abbreviation". For example, Hamas and Fatah (the main political parties in Palestine) are both acronyms. Fatah is actually a reverse acronym.
We're defending the rights of the oppressed minority against the Jews, and that oppressed minority that unites under the party "conquest" ? Seriously ?
If this is your name, how the hell would you defend the idea that palestinians are just trying to live independantly ?
I think that's not strictly true, what about الخ meaning الى اّخره (i.e. to the end of it, meaning "etc".). In some books ح can mean حديث. But generally it is very rare.
Arabic hasn't "modernised" in quite the same way many western languages have, which makes it quite interesting. Since it has stayed a written-only language for quite some time, the letters were not broken up, and its joined-up form has stayed into the digital age (at least, I assume this is why)
I don't mean just abbreviation, I mean Arabic hasn't seen certain types of changes that have occurred in Western languages. Although, arguably, Latin has always been fundamentally a non-cursive script, and Arabic has always been cursive, perhaps.
Ah, I see. I think you may have it backwards, though: script Arabic developed around the same time as lower-case letters in Greek and Latin (very roughly 0CE), coinciding one presumes with growing use of paper for writing. While Latin eventually incorporated its older "upper-case" forms with a minor grammatical function, Arabic discarded them altogether as antiquated. In that sense, one might well make the case that Arabic is in fact the more "modern" alphabet.
You're quite correct. The old 'Latin' script (our uppercase letters) where revived in the renaissance - and were slowly merged with the Hunnish script (our lower case letters) - about the same time as they adopted Arabic numerals (read right to left, for added confusion).
Other alphabets (Greek and Cyrillic) created capital letters in emulation much later (circa 1800's). Most alphabets/sylabaries (Arabic, Hebrew, Korean, Japanese, etc) don't have upper and lower cases.
Likewise punctuation, which prosidic languages (like English) need to signal stress and emphasis which can transform meaning, but which are also absent from lots of other writing systems.
I think talk of 'more modern' or 'less modern' is nonsense though. Writing is an imperfect representation of speaking, and whatever works is just fine.
I don't think the timeline for the development of the Arabic script is quite right. I believe it's more typically put 3-400 CE, with the Nabatean script slowly accruing Arabic-like features over time.
Of course, my knowledge of the history of writing extends to a single undergrad course and a quick skim of Wikipedia as a refresher. I'll happily admit I'm wrong if that turns out to be the case.
I didn't mean to imply modern Arabic script came to be then -- it wasn't what we'd call Arabic today before the Quran -- only that evolution from inscribed block capitals to written script was happening at around the same time as that transition in other languages.
Anyway, in this context I'm counting 300BCE - 300CE as "very roughly 0CE".
Even so it still seems apparent that modern Latin alphabets have been adapted for typesetting, whereas even contemporary written Arabic seems to remain optimized for handwriting. It seems reading Arabic is a lot more difficult than reading Latin based writing, language issues aside, simply becuase the letters are all joined together.
I don't see how cursive script makes Arabic less “modern.” It might make carving it onto stone a bit harder, but the printing press, the typewriter, and the computer all had no problems producing quality Arabic type.
Having written code which renders Arabic text, I found “joining” up the letters to be quite simple. It's just a few rules to choose which glyph to display for each character depending on context. The tricky bits, I found, is in integrating a right-to-left script with systems which were made with only left-to-right in mind.
Ever tried dealing with vertical column text (e.g. Japanese?). I haven't myself, but the fact Microsoft made fonts on Japanese versions of Windows have special "@"-prefixed versions rotated 90 degrees, so if a document is written in it, then changed to that font and rotated, it's the right way up, makes me suspect it's pretty difficult.
That sounds like the result of a lack of forethought when designing the font rendering code, and not because it's a big problem. I've used vertical Japanese text input extensively in Word without any problems.
mixing RTL and LTR is a pain in unix-like tools (gnu, os x).
What I usually (not always) found out diring my experiments woth Persian is that the gnu tools themselves deal with it quite fine - but the terminal application shows them at pretty random places. If there is both LTR and RTL in one file, havoc usually ensues.
(persian also has this funny "invisible space", but it causes only small annoyances)
I digress, but I think this is one of the most annoying things about Unix these days. I practically live by the command line; but terminals are terrible. We're basically emulating technology that was already getting obsolete in the 70s!
Plan 9 had the courage to shed this cruft by having simple text windows which have a prompt. No curser addressing and no crazy control codes. That makes rendering it just as easy (and beautiful) as rendering text box contents. I wish the Linux desktop environments would follow suit.
There was a company that applied to YC that wanted to do the "Google Docs for Statistics" but was rejected. They wrote about it in a blog post but I can't find it. They ended up not launching.
It will be worth it to connect with these people to see if there is anything that can be learned from them.
I believe the company talbina's thinking of is mine: We applied as Theoryville, got interviewed, got rejected, applied to Betaspring (http://betaspring.com/), got accepted, changed our name to DataBraid, and proceeded to fall apart over the course of a summer.
I do think the idea has a lot of potential, and what StatWing has built is already more complete than what my team managed to build in 3 months. A few suggestions I'd offer based on that experience:
1. Parsing CSVs is easy in theory, but painful in practice, because CSVs in the wild tend to be full of junk. I would provide a JSON API that makes it easy for developers to put data in your system directly, allowing people to build their own CSV parsers for you.
2. Use GitHub as your model. You want people to collaborate around data the same way that developers collaborate around code. Just about every day when I was doing DataBraid, we'd discuss a use case and then say "Oh, GitHub already figured out the right way to do this." The most compelling use case here is that researchers can run different sets of tests on the same data and discuss which approach is the most valid/insightful.
3. Getting to revenue will be hard, but having paying customers will make it much, much easier to attract investment. So find the MVP that people will pay for and put everything else on a "nice-to-have" list.
It allowed markup inside the pull down so that when you hover on an icon in the pull down list, a tooltip appeared. It was much more, but this is what I remember.
I am having a hard time validating the idea so I decided to post it here and take a little embarrassment in the hopes of getting feedback. Excuse the domain name and the eye soar design.
Are there any other decision tables you need? I am willing to make a decision table for anyone (will take me a few days) in the hopes of getting feedback as well.
Heroku (YC08) has been around longer than #22 (http://ycombinator.com/ideas.html), but pg's item on fundable ideas might hint at some of his vision regarding abstracted platforms:
Don't make it feel like a database. That frightens people. The question to ask is: how much can I let people do without defining structure? You want the database equivalent of a language that makes its easy to keep data in linked lists.
I realize that Derby also helps build real-time applications, but I haven't heard anything about Derby being in YC. Has there been a public announcement that they are?
If we're just listing frameworks that help build real-time systems, Pusher would be another example.
Samuel, I just wanted to let you know that you are a huge inspiration to me even though you don't know me. I remember your first post when you said you did the first version of your site while commuting on the NYC subway every day.
That small piece of info is my driving force every week as I recall how productive other entrepreneurs can be. So thank you.
Originally this blog post was going to be a "and so can you" type post. I thought it was kind of important to set the tone to say that I'm a schlub who just pushed hard enough on the persistence/patience/perseverance/perspiration mix. It ended up being all about me instead, but I'm still working on the post in my head.
I'm not especially productive. I just work relentlessly on NewsBlur, every single day. Check the graph on http://github.com/samuelclay. There's another year before that of the same persistent pushing. And oftentimes I think that I take more than twice the time others do to get something like a distributed feed fetcher built. But push through it and one day you're shipping.
My first year was all internal motivation. But after I launched, that very quickly switched to external motivation. I added NewsBlur's Get Satisfaction forum to the dashboard of every NewsBlur user. That way I get a lot more feedback about the product. I had to build that little table by hand, even writing a cronjob to grab the latest feedback so it would show up instantly for users just logging in, but that was so clearly the basis for my external motivation.
Getting paid, on the other hand, was never a motivation. It just didn't occur to me that I could live off NewsBlur. Even today, that's not the case. But I'm taking a gamble by going full-time, spending the next few months cranking out my big social branch, and seeing if there's something bigger in NewsBlur. Luckily, at this rate, simply monetizing the current site might be enough to keep me off the streets.
I had a similar experience to you 2004-2007 with an RSS-to-everything service called Feed Digest. It started as a way to get my Delicious links onto my blog and eventually turned into my full time job with 25k users. I sold it in 2007 after a number of mistakes on my part: http://peterc.org/blog/2010/257-three-years-ago-i-sold-my-st...
Anyway, the only reason I'm piping up is because I think NewsBlur is cool and I hope you manage to avoid the single bigger killer for my own project.. not charging enough! :-) I earned enough for it to be an OK living but not enough to hire people, get much help, etc, and it turned into a real burnout situation. There wasn't anything like Hacker News to learn from at the time either.
So if you're going full time with this, I wish you the very best of luck but I pray (and I'm not a religious man!) you find a scalable and realistic business model that makes this project comfortable to run, rather than just enough to keep you off the streets, as it were.
The original idea started in 2006 while I was working with Eight Media, which I had founded together with Daniel Spronk (spazmaster). After hacking away at a prototype for a while we presented our MVP to a live audience in London in 2007. The response was overwhelming, we got serious about it. Found an angel willing to invest.
Then in 2009 Daniel went back to focus on Eight.nl while I kept going with Soocial. Looking back I think it was much harder without Daniel as a co-founder than I expected. It makes such a big difference to have someone there to bounce idea's, make decisions together and just have the support of someone else who's just as involved and committed.
Very much agree with PG's statement that although possible as a single founder, you're chances of success are less.
