More information

Questions: What’s in the package? | What do I need to run ICECUP? | What can ICECUP do? | Will ICECUP continue to be developed? | Feedback

The DCPSE Sample Corpus is available for download NOW.

It comes complete with 7 texts selected by Gerry Nelson from the ICE-GB Corpus, and the state-of-the-art ICECUP 3.1 software written by Sean Wallis.

WHAT IS IN THE SAMPLE CORPUS PACKAGE? 

  • Five texts from the ICE-GB part of the corpus (over 10,000 words) plus two texts from the LLC part (another 10,000 plus words), fully parsed and annotated. These are exactly as they are in DCPSE.
  • The latest release of ICECUP 3.1. This is a full working version of the software (see below) complete with help.
  • Example Fuzzy Tree Fragments.

The sample contains the following ten texts, shown in the last column. You can view these texts and their classification when you download and install the software. The complete DCPSE structure is visible from ICECUP’s Corpus Map.

Note that the LLC texts are 5,000-word samples, whilst the ICE texts each contain 2,000 words. The entire corpus is balanced as 200x2,000 words (ICE) + 80x5,000 words (LLC) = 800,000 words.

Face-to-face conversations (154) Formal (28) DI-A13 DL-A01
Informal (126) DI-B29
DI-B82
DL-B04
Telephone conversations (14)  
Broadcast discussions (28) DI-D13
Broadcast interviews (14)  
Spontaneous commentary (32) DI-F15
Parliamentary language (7)  
Legal cross-examination (3)  
Assorted spontaneous (7)  
Prepared speech (21)  

WHAT IS NOT INCLUDED?

DCPSE is supplied on CD-ROM. DCPSE contains two hundred and eighty texts of spoken British English from across the decades. To obtain the other 273 texts, you must order the CD-ROM. If you want to do this, click here.

WHAT DO I NEED TO RUN ICECUP 3.1? 

ICECUP runs on PCs under Windows 3.1 and above. It has been tested exhaustively on Windows 3.1, 95+ and XP. Owing to the nature of the program, we recommend a fast processor and a fast hard disk, although these are not essential. ICECUP will run on any stand-alone or networked PC from a 386 running Windows 3.1 with 32MB upwards.

Sampler system requirements

There are two install packages, with hard disk capacity requirements as follows:

Process Text
To download: 3.5Mb
To install: 21Mb
To run: 8.1Mb

We have tested the software extensively on platforms that we have access to, and we have witnessed it working OK on others. The most up-to-date list is shown below.

Software platform Status
Windows 3.1/WfW3.11 Regularly tested
Windows 95/98/ME Heavily tested, in constant use
Windows NT4/WTS/2000/XP Regularly tested, in frequent use
OS/2 Not tested
Mac/PPC Windows emulator Tested OK

Feedback: Please tell us if you try to run ICECUP on platforms other than PCs running Windows 95/98/ME or Windows 3.1. We want to know about your software problems, to help you and other end users. Since we are supplying software free, and “as is” (see the licence agreement) with a Sample Corpus taken from ICE-GB, we think that it is only fair that you give us some feedback.

If ICECUP runs brilliantly on Windows XP or crashes dismally on a Mac, tell us. Email us at the Survey (s.wallis@ucl.ac.uk) so that (a) we can try to solve any outstanding problems, and (b) tell others about them.

System requirements for the DCPSE corpus (CD-ROM)

These differ from the above only in terms of hard disk space. You need 90Mb to install the entire corpus. Note that you can also run searches off the CD without installing anything.

The software is identical to that supplied with the sample corpus. You can therefore ‘try before you buy’. If in doubt, install the sample corpus and software before ordering the CD.

email: s.wallis@ucl.ac.uk


WILL ICECUP CONTINUE TO BE DEVELOPED?

Yes. ICECUP 3.1 is simply the latest release of the software. The Next Generation Tools project will develop more software primarily focused on linguistic experimentation, on this platform.


COMMENTS, SUGGESTIONS, QUERIES...

The ICECUP 3.1 help file provides extensive guidance on installing ICECUP and the corpus. It includes a "How do I?" section addressing many common questions and a complete grammar reference.

The Resources section of this website has sections on Fuzzy Tree Fragments and an explanation of carrying out experiments with parsed corpora.

If you want to ask a question of the author you can email Sean Wallis directly. To join the ICECUP beta programme click here.

This page last modified 12 June, 2013 by Survey Web Administrator.