Before the invention of computers, all bundles were created manually by photocopying the documents to be included in the Bundle, arranging them in order, page-numbering them using a black ball point pen, typing up and printing the Index (which lists every document with its page number) and then photocopying the result as many times as necessary to produce the required number of copies of the Bundle.
But there are a number of problems with doing it manually. First of all creating a Bundle manually takes some time, and if a bundle needs to be changed after it has been produced (e.g. by inserting new documents, or rearranging the documents into a more useful and logical order) this also can take a lot of time, not least because every change has to be done to every copy which has been produced.
The production of the bundle is a process which the parties must collaborate in. One party will be responsible for actually producing it in the required number of copies but the other party must co-operate in confirming which documents they wish to be included and it raising promptly any objection to the inclusion of documents in those rare circumstances where this is appropriate (e.g. If the other party is seeking to include a without prejudice offer document). This means that the bundle producing party has to start off by proposing a set of documents which the bundle is to contain and seek the other party's comments. But how exactly is that to be done? Because of the work involved in producing a bundle, and the extra work involved in changing it after it has been produced, the option of producing a draft bundle, sending it to the other party for their comments, and then producing a further draft, is unattractive.
What will typically be done is that the bundle producing party will produce, initially, not a full draft bundle but a draft index and send that to the other party for comments. But often the draft index is not sufficiently detailed for the other party to be sure what documents the bundle producing party is proposing to be included. This is because a draft index is based on the sort of index which would typically be used for a full bundle. Because typing up the index for a bundle is a laborious task, certain shortcuts are typically employed. For example, depending on the case, often a fair amount correspondence between the parties after commencement of litigation will be included in the bundle which is considered to be needed but not thought to be likely to be referred to at the hearing very often, so the index might simply say "Selected Litigation Correspondence" and just give the page number where the correspondence starts without listing each item of that correspondence. Strictly speaking failing to list all documents is contrary to the rules (at least in court cases subject to CPR PD 39A para. 3.5) but it is often done as a shortcut, when a bundle is being produced manually, and if done intelligently may cause no real problems at the hearing when the full bundle (index and paginated documents) is used. However it is obvious that such a draft index (unaccompanied by the paginated documents which will be in the bundle) cannot identify with certainty the documents which a party is proposing to include. If you receive a draft index which lists some documents individually but does not list others, simply referring to them as "Selected Litigation Correspondence" you cannot tell what correspondence that is intended to refer to. A similar problem arises where a document is listed individually but the description is not clear: if you had the complete bundle you could easily refer to the document itself by page number to double-check which document it is in case of doubt, but if you just have an initial draft index you cannot do that.
Even between parties who genuinely want to be businesslike and co-operate in the production of the hearing bundle, the fallibility of the "agree by draft index" method, coupled with the desire not to start the bundle production process too early, can, at worst, lead to increasingly frantic, complex and critical emails between the parties, to late production of the bundle omitting documents one party wishes to have included, and to the consequent last minute production by the other party of a supplementary bundle (which is not ideal).
That can happen with parties who genuinely wish to co-operate but if one party indulges in gamesmanship it can be worse with much time and energy wasted and the potential for the first part of the hearing being taken up with arguments about whose fault it was, to the detriment of both parties in terms of time and cost.
To avoid all these problems it is much better to use computer software such as Caselines. Using Caselines, the documents which are candidates for inclusion in the bundle are loaded as PDFs and given a date and brief description, and the system automatically produces an indexed and paginated bundle in PDF form. If documents are added to Caselines then the PDF bundle is automatically changed. Likewise if documents are removed again the PDF bundle is automatically changed. As well as making the process much easier for everyone, a further advantage is that there is no need for the bundle production process to be left to the last minute: it can start several weeks before the hearing and both parties can see the full bundle - actual documents as well as index - as it is refined in the light of each side's comments.
Usually if there is clarity about which documents each side is referring to there are no real disputes about what the bundle should contain, but if there should be some live dispute - for example if one party claims that a document is without prejudice and should not be included while the other argues that it was not without prejudice - if the bundle production process is started in good time - and there is no reason why it should not be if computer software (rather than the manual system) is used - it is normally possible for such disputes to be resolved thus saving time at the hearing.