Path: csiph.com!x330-a1.tempe.blueboxinc.net!newsfeed.hal-mli.net!feeder3.hal-mli.net!nx01.iad01.newshosting.com!newshosting.com!news2.euro.net!feeder.news-service.com!eternal-september.org!feeder.eternal-september.org!.POSTED!not-for-mail
From: Joshua Cranmer <Pidgeot18@verizon.invalid>
Newsgroups: comp.lang.java.programmer
Subject: Re: higher precision doubles
Date: Mon, 08 Aug 2011 21:34:23 -0500
Organization: A noiseless patient Spider
Lines: 33
Message-ID: <j1q6bm$3cp$1@dont-email.me>
References: <j1hqc1$1ch$1@news.albasani.net> <strictfp-20110806003831@ram.dialup.fu-berlin.de> <j1k4db$1n2$1@news.albasani.net> <j1k87t$kol$1@dont-email.me> <j1kb73$h3p$1@news.albasani.net> <j1kg3g$61c$1@dont-email.me> <j1kheh$ta4$1@news.albasani.net> <GMCdnfvkB6IaUaDTnZ2dnUVZ_omdnZ2d@earthlink.com> <j1kj1j$be$1@news.albasani.net> <t8idnWDwz_v8iKPTnZ2dnUVZ_gCdnZ2d@earthlink.com> <j1manb$2lu$1@news.albasani.net> <nOednS4yGv5WIaPTnZ2dnUVZ_gqdnZ2d@earthlink.com> <j1momo$34q$1@news.albasani.net> <j1nbr7$h7r$1@dont-email.me> <j1o1sk$427$1@news.albasani.net> <DbednYquHMbkRaLTnZ2dnUVZ_s2dnZ2d@earthlink.com> <j1p56f$kfq$1@news.albasani.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Tue, 9 Aug 2011 02:34:30 +0000 (UTC)
Injection-Info: mx04.eternal-september.org; posting-host="B3q1fNdvNsCxx/IZ4idKGA"; logging-data="3481"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19jZXCeyKYomMlmVcBNsljUc5lcDAypIS8="
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:5.0) Gecko/20110624 Thunderbird/5.0
In-Reply-To: <j1p56f$kfq$1@news.albasani.net>
Cancel-Lock: sha1:1/lTt6/LaQsAd3nuF4f3O7+PeeM=
Xref: x330-a1.tempe.blueboxinc.net comp.lang.java.programmer:6893

On 8/8/2011 12:08 PM, Jan Burse wrote:
> Just assume I want to put the functionality into an Applet. Means
> somehow the aspect of supporting a couple of plattforms and
> architectures is important.

Using native APIs in an applet is impossible without signing your 
applet. This would either imply using BigDecimal, signing your code, or 
making do with a pure software implementation.

Also note that not all architectures support an 80-bit 
extended-precision floating point format. Not that ARM sports an FPU, 
but its most common FPU coprocessor fails to implement 80-bit width, and 
I should mention that ARM processors are by far much more common than 
the install base of x86. In fact, I think the only processor with a 
significant install base that has hardware support for a floating point 
number larger than 64 bits is the x86.

> Also one could implement mutable objects (instances) instead of
> value objects (instances). So that instead of an API of the form:

Small mutable objects like this are likely to negatively impact GC.

> One could also do:
>
> myMath:
> sin(myMutableObjectClass x, myMutableObjectClass res);

If you're only using the mutable object for the return value, you are 
probably taking a very small performance hit.

-- 
Beware of bugs in the above code; I have only proved it correct, not 
tried it. -- Donald E. Knuth