谢谢，我的朋友

I think it is not something about friendly or friendlily or write for that matter, it’s about make. Make can do a lot of things that other verbs cannot, one of which is to take a small clause, patterning with consider: “I made it friendly”, “I consider it friendly”. I believe “I made it friendly” can be paraphrased as “I made it to be friendly” without changing the meaning like the one with write.

This small clause, however, cannot be pulled off by verbs like write. Just like you cannot say “I wrote it friendly to the non native speaker of English.” but you can say “I made it friendly to non native speaker of English”.

Then your first two questions can be answered if we say verb + object + adj is a small clause structure which make allows but not write. A further question would be why. It seems to link to Pinker’s solution to what is called Baker’s Paradox in acquisition. (Pinker 1989) He argues that it’s the narrow conflation classes of lexical semantics that help children to know “toss me the paper” is ok but “push me the paper”, so argument structure is sensitive to subtle semantics. The verb write seems to emphasize the means to create something but make is more generic. Evidence of this would be the fact that “I made it friendly to non English native speaker” can mean I wrote something or created something by other means.

About your last question, words like something is famous in taking post-modifier and forbidding pre-modifier. Pronouns like it cannot do this, as are most nominals. In general, pronouns don’t take adjectival modifiers except in “the Despicable Me”, much less post-modifiers. And I don’t know anything about the property of something . One possible way to look at would be that something is from some + thing. With some occupying the determiner position, there is no position in between, as you cannot add adjective to the left of determiner “*beautiful my mother” and “my beautiful mother”. The question would be why does something allows to have modifier in the final position but not it.

Hope this helps.

Zheng

The first line of the SQL Developer output is always:

REM INSERTING into tableName

…so we can get the name of the table very easily, without resort to our crazy regular expression. Moreover, we don’t need to convert &s to their ASCII equivalent, we just use the set define off; declaration at the beginning of the script. Combined, this greatly simplifies things:

gg$yiw:%s/"\(<C-r>"\)"/\1/g
ggOtruncate table <Esc>pA;<Esc>Oset define off;
<Esc>:%s/^REM.*$//g
Go
commit;

ubyte q[]=[59, 105, 109, 112, 111, 114, 116, 32, 115,
116, 100, 46, 115, 116, 100, 105, 111, 59, 118, 111,
105, 100, 32, 109, 97, 105, 110, 40, 41, 123, 119, 114,
105, 116, 101, 40, 34, 117, 98, 121, 116, 101, 32, 113,
91, 93, 61, 34, 44, 113, 44, 99, 97, 115, 116, 40, 115,
116, 114, 105, 110, 103, 41, 113, 41, 59, 125, 10];
import std.stdio;void main(){write("ubyte q[]=",q,
cast(string)q);}

(Again, Unix EOL and ignore the line wrapping in the display, here.)

import std.stdio;

immutable ubyte quine[] = [
0x0a,0x5d,0x3b,0x0a,0x0a,0x76,0x6f,0x69,0x64,0x20,0x6d,0x61,0x69,0x6e,0x28,0x29,0x20,0x7b,0x0a,0x20,0x20,0x77,0x72,0x69,0x74,0x65,0x6c,0x6e,0x28,0x22,0x69,0x6d,0x70,0x6f,0x72,0x74,0x20,0x73,0x74,0x64,0x2e,0x73,0x74,0x64,0x69,0x6f,0x3b,0x22,0x2c,0x0a,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x22,0x5c,0x6e,0x5c,0x6e,0x22,0x2c,0x0a,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x20,0x22,0x69,0x6d,0x6d,0x75,0x74,0x61,0x62,0x6c,0x65,0x20,0x75,0x62,0x79,0x74,0x65,0x20,0x71,0x75,0x69,0x6e,0x65,0x5b,0x5d,0x20,0x3d,0x20,0x5b,0x22,0x29,0x3b,0x0a,0x20,0x20,0x66,0x6f,0x72,0x65,0x61,0x63,0x68,0x20,0x28,0x71,0x3b,0x20,0x71,0x75,0x69,0x6e,0x65,0x29,0x20,0x77,0x72,0x69,0x74,0x65,0x66,0x28,0x22,0x30,0x78,0x25,0x30,0x32,0x78,0x2c,0x22,0x2c,0x71,0x29,0x3b,0x0a,0x20,0x20,0x66,0x6f,0x72,0x65,0x61,0x63,0x68,0x20,0x28,0x71,0x3b,0x20,0x71,0x75,0x69,0x6e,0x65,0x29,0x20,0x77,0x72,0x69,0x74,0x65,0x28,0x63,0x61,0x73,0x74,0x28,0x63,0x68,0x61,0x72,0x29,0x20,0x71,0x29,0x3b,0x0a,0x7d,0x0a,
];

void main() {
  writeln("import std.stdio;",
          "\n\n",
          "immutable ubyte quine[] = [");
  foreach (q; quine) writef("0x%02x,",q);
  foreach (q; quine) write(cast(char) q);
}

While it works in the same way — and despite the hexdump — it’s a bit easier to understand. Basically, the array stores the binary data of the source code (except the header and the data itself), then the main function just spits it all out: first as the hexdump and then encoded as characters (note the Unix line endings: 0x0a). It’s a bit more D, in style, with the foreach statements…but I can still do better!